ae.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Imputting data using autoencoder\n",
    "Based on [this code](https://github.com/lstruski/Processing-of-missing-data-by-neural-networks/blob/master/ae.py)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "is_executing": false,
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Read data done in 1.118s.\n",
      "GMM done in 5.849s.\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "from datetime import datetime\n",
    "from time import time\n",
    "\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from PIL import Image\n",
    "from sklearn.impute import SimpleImputer\n",
    "from sklearn.mixture import GaussianMixture\n",
    "from tqdm import tqdm\n",
    "\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'\n",
    "\n",
    "RANDOM_SEED = 42\n",
    "tf.set_random_seed(RANDOM_SEED)\n",
    "\n",
    "# Training Parameters\n",
    "learning_rate = 0.01\n",
    "n_epochs = 250\n",
    "batch_size = 64\n",
    "# ds_label: str = \"Mnist\"\n",
    "ds_label: str = \"Fashion\"\n",
    "    \n",
    "\n",
    "# Network Parameters\n",
    "num_hidden_1 = 256  # 1st layer num features\n",
    "num_hidden_2 = 128  # 2nd layer num features (the latent dim)\n",
    "num_hidden_3 = 64  # 3nd layer num features (the latent dim)\n",
    "num_input = 784  # MNIST data_rbfn input (img shape: 28*28)\n",
    "\n",
    "n_distribution = 5  # number of n_distribution\n",
    "\n",
    "width_mask = 13  # size of window mask\n",
    "\n",
    "# tf Graph input (only pictures)\n",
    "X = tf.placeholder(\"float\", [None, num_input])\n",
    "\n",
    "initializer = tf.contrib.layers.variance_scaling_initializer()\n",
    "\n",
    "weights = {\n",
    "    'encoder_h1': tf.Variable(initializer([num_input, num_hidden_1])),\n",
    "    'encoder_h2': tf.Variable(initializer([num_hidden_1, num_hidden_2])),\n",
    "    'encoder_h3': tf.Variable(initializer([num_hidden_2, num_hidden_3])),\n",
    "    'decoder_h1': tf.Variable(initializer([num_hidden_3, num_hidden_2])),\n",
    "    'decoder_h2': tf.Variable(initializer([num_hidden_2, num_hidden_1])),\n",
    "    'decoder_h3': tf.Variable(initializer([num_hidden_1, num_input])),\n",
    "}\n",
    "biases = {\n",
    "    'encoder_b1': tf.Variable(tf.random_normal([num_hidden_1])),\n",
    "    'encoder_b2': tf.Variable(tf.random_normal([num_hidden_2])),\n",
    "    'encoder_b3': tf.Variable(tf.random_normal([num_hidden_3])),\n",
    "    'decoder_b1': tf.Variable(tf.random_normal([num_hidden_2])),\n",
    "    'decoder_b2': tf.Variable(tf.random_normal([num_hidden_1])),\n",
    "    'decoder_b3': tf.Variable(tf.random_normal([num_input])),\n",
    "}\n",
    "\n",
    "\n",
    "def generate_mask(x, y, size):\n",
    "    return np.concatenate([28 * i + np.arange(x, x + size) for i in\n",
    "                           np.arange(y, y + size)], axis=0).astype(np.int32)\n",
    "\n",
    "\n",
    "def data_with_defined_mask(x, mask):\n",
    "    for i in range(x.shape[0]):\n",
    "        x[i, generate_mask(mask[i][0], mask[i][1], 12)] = np.nan\n",
    "    return x\n",
    "\n",
    "\n",
    "def nr(mu, sigma):\n",
    "    non_zero = tf.not_equal(sigma, 0.)\n",
    "    new_sigma = tf.where(non_zero, sigma, tf.fill(tf.shape(sigma), 1e-20))\n",
    "    sqrt_sigma = tf.sqrt(new_sigma)\n",
    "\n",
    "    w = tf.div(mu, sqrt_sigma)\n",
    "    nr_values = sqrt_sigma * (tf.div(tf.exp(tf.div(-tf.square(w), 2.)), np.sqrt(2 * np.pi)) +\n",
    "                              tf.multiply(tf.div(w, 2.), 1 + tf.erf(tf.div(w, np.sqrt(2)))))\n",
    "\n",
    "    nr_values = tf.where(non_zero, nr_values, (mu + tf.abs(mu)) / 2.)\n",
    "    return nr_values\n",
    "\n",
    "\n",
    "def conv_first(x, means, covs, p, gamma):\n",
    "    gamma_ = tf.abs(gamma)\n",
    "    # gamma_ = tf.cond(tf.less(gamma_[0], 1.), lambda: gamma_, lambda: tf.square(gamma_))\n",
    "    covs_ = tf.abs(covs)\n",
    "    p_ = tf.nn.softmax(p, axis=0)\n",
    "\n",
    "    check_isnan = tf.is_nan(x)\n",
    "    check_isnan = tf.reduce_sum(tf.cast(check_isnan, tf.int32), 1)\n",
    "\n",
    "    x_miss = tf.gather(x, tf.reshape(tf.where(check_isnan > 0), [-1]))  # data_rbfn with missing values\n",
    "    x = tf.gather(x, tf.reshape(tf.where(tf.equal(check_isnan, 0)), [-1]))  # data_rbfn without missing values\n",
    "\n",
    "    # data_rbfn without missing\n",
    "    layer_1 = tf.nn.relu(tf.add(tf.matmul(x, weights['encoder_h1']), biases['encoder_b1']))\n",
    "\n",
    "    # data_rbfn with missing\n",
    "    where_isnan = tf.is_nan(x_miss)\n",
    "    where_isfinite = tf.is_finite(x_miss)\n",
    "    size = tf.shape(x_miss)\n",
    "\n",
    "    weights2 = tf.square(weights['encoder_h1'])\n",
    "\n",
    "    # Collect distributions\n",
    "    distributions = tf.TensorArray(dtype=x.dtype, size=n_distribution)\n",
    "    q_collector = tf.TensorArray(dtype=x.dtype, size=n_distribution)\n",
    "\n",
    "    # Each loop iteration calculates all per component\n",
    "    def calculate_component(i, collect1, collect2):\n",
    "        data_miss = tf.where(where_isnan, tf.reshape(tf.tile(means[i, :], [size[0]]), [-1, size[1]]), x_miss)\n",
    "        miss_cov = tf.where(where_isnan, tf.reshape(tf.tile(covs_[i, :], [size[0]]), [-1, size[1]]),\n",
    "                            tf.zeros([size[0], size[1]]))\n",
    "\n",
    "        layer_1_m = tf.add(tf.matmul(data_miss, weights['encoder_h1']), biases['encoder_b1'])\n",
    "        layer_1_m = nr(layer_1_m, tf.matmul(miss_cov, weights2))\n",
    "\n",
    "        norm = tf.subtract(data_miss, means[i, :])\n",
    "        norm = tf.square(norm)\n",
    "        q = tf.where(where_isfinite,\n",
    "                     tf.reshape(tf.tile(tf.add(gamma_, covs_[i, :]), [size[0]]), [-1, size[1]]),\n",
    "                     tf.ones_like(x_miss))\n",
    "        norm = tf.div(norm, q)\n",
    "        norm = tf.reduce_sum(norm, axis=1)\n",
    "\n",
    "        q = tf.log(q)\n",
    "        q = tf.reduce_sum(q, axis=1)\n",
    "\n",
    "        q = tf.add(q, norm)\n",
    "\n",
    "        norm = tf.cast(tf.reduce_sum(tf.cast(where_isfinite, tf.int32), axis=1), tf.float32)\n",
    "        norm = tf.multiply(norm, tf.log(2 * np.pi))\n",
    "\n",
    "        q = tf.add(q, norm)\n",
    "        q = -0.5 * q\n",
    "\n",
    "        return i + 1, collect1.write(i, layer_1_m), collect2.write(i, q)\n",
    "\n",
    "    i = tf.constant(0)\n",
    "    _, final_distributions, final_q = tf.while_loop(lambda i, c1, c2: i < n_distribution, calculate_component,\n",
    "                                                    loop_vars=(i, distributions, q_collector),\n",
    "                                                    swap_memory=True, parallel_iterations=1)\n",
    "\n",
    "    distrib = final_distributions.stack()\n",
    "    log_q = final_q.stack()\n",
    "\n",
    "    log_q = tf.add(log_q, tf.log(p_))\n",
    "    r = tf.nn.softmax(log_q, axis=0)\n",
    "\n",
    "    layer_1_miss = tf.multiply(distrib, r[:, :, tf.newaxis])\n",
    "    layer_1_miss = tf.reduce_sum(layer_1_miss, axis=0)\n",
    "\n",
    "    # join layer for data_rbfn with missing values with layer for data_rbfn without missing values\n",
    "    layer_1 = tf.concat((layer_1, layer_1_miss), axis=0)\n",
    "    return layer_1\n",
    "\n",
    "\n",
    "# Building the encoder\n",
    "def encoder(x, means, covs, p, gamma):\n",
    "    layer_1 = conv_first(x, means, covs, p, gamma)\n",
    "\n",
    "    # Encoder Hidden layer with sigmoid activation\n",
    "    layer_2 = tf.nn.sigmoid(tf.add(tf.matmul(layer_1, weights['encoder_h2']), biases['encoder_b2']))\n",
    "    layer_3 = tf.nn.sigmoid(tf.add(tf.matmul(layer_2, weights['encoder_h3']), biases['encoder_b3']))\n",
    "    return layer_3\n",
    "\n",
    "\n",
    "# Building the decoder\n",
    "def decoder(x):\n",
    "    layer_1 = tf.nn.sigmoid(tf.add(tf.matmul(x, weights['decoder_h1']), biases['decoder_b1']))\n",
    "    layer_2 = tf.nn.sigmoid(tf.add(tf.matmul(layer_1, weights['decoder_h2']), biases['decoder_b2']))\n",
    "    layer_3 = tf.nn.sigmoid(tf.add(tf.matmul(layer_2, weights['decoder_h3']), biases['decoder_b3']))\n",
    "    return layer_3\n",
    "\n",
    "\n",
    "def prep_x(x):\n",
    "    check_isnan = tf.is_nan(x)\n",
    "    check_isnan = tf.reduce_sum(tf.cast(check_isnan, tf.int32), 1)\n",
    "\n",
    "    x_miss = tf.gather(x, tf.reshape(tf.where(check_isnan > 0), [-1]))\n",
    "    x = tf.gather(x, tf.reshape(tf.where(tf.equal(check_isnan, 0)), [-1]))\n",
    "    return tf.concat((x, x_miss), axis=0)\n",
    "\n",
    "\n",
    "t0 = time()\n",
    "images = list()\n",
    "for i in range(10_000):\n",
    "    img: Image = Image.open('{}/img/{}.jpg'.format(ds_label,i))\n",
    "    images.append(np.array(img))\n",
    "images = np.array(images)\n",
    "labels = list()\n",
    "masks = list()\n",
    "labels_file = open('labels{}.csv'.format(ds_label), 'r')\n",
    "for line in labels_file:\n",
    "    line_list: list = line.split(',')\n",
    "    labels.append(int(line_list[1]))\n",
    "    masks.append((int(line_list[2]), int(line_list[3])))\n",
    "labels = np.array(labels)\n",
    "\n",
    "x_train, y_train = images[:8000] / 255.0, labels[:8000]\n",
    "x_test, y_test = images[8000:] / 255.0, labels[8000:]\n",
    "print(\"Read data done in %0.3fs.\" % (time() - t0))\n",
    "\n",
    "data_train = x_train.reshape(-1, num_input)\n",
    "data_test = x_test.reshape(-1, num_input)\n",
    "\n",
    "data_train = data_with_defined_mask(data_train, masks)\n",
    "\n",
    "data_test = data_with_defined_mask(data_test, masks[8000:])\n",
    "\n",
    "imp = SimpleImputer(missing_values=np.nan, strategy='mean')\n",
    "data = imp.fit_transform(data_train)\n",
    "\n",
    "t0 = time()\n",
    "gmm = GaussianMixture(n_components=n_distribution, covariance_type='diag').fit(data)\n",
    "print(\"GMM done in %0.3fs.\" % (time() - t0))\n",
    "\n",
    "p = tf.Variable(initial_value=np.log(gmm.weights_.reshape((-1, 1))), dtype=tf.float32)\n",
    "means = tf.Variable(initial_value=gmm.means_, dtype=tf.float32)\n",
    "covs = tf.Variable(initial_value=gmm.covariances_, dtype=tf.float32)\n",
    "gamma = tf.Variable(initial_value=tf.random_normal(shape=(1,), mean=1., stddev=1.), dtype=tf.float32)\n",
    "del data, gmm\n",
    "\n",
    "# Construct model\n",
    "encoder_op = encoder(X, means, covs, p, gamma)\n",
    "decoder_op = decoder(encoder_op)\n",
    "\n",
    "y_pred = decoder_op  # prediction\n",
    "y_true = prep_x(X)  # Targets (Labels) are the input data_rbfn.\n",
    "\n",
    "where_isnan = tf.is_nan(y_true)\n",
    "y_pred = tf.where(where_isnan, tf.zeros_like(y_pred), y_pred)\n",
    "y_true = tf.where(where_isnan, tf.zeros_like(y_true), y_true)\n",
    "\n",
    "# Define loss and optimizer, minimize the squared error\n",
    "loss = tf.reduce_mean(tf.pow(y_true - y_pred, 2))\n",
    "optimizer = tf.train.RMSPropOptimizer(learning_rate).minimize(loss)\n",
    "\n",
    "# Initialize the variables (i.e. assign their default value)\n",
    "init = tf.global_variables_initializer()\n",
    "\n",
    "trn_summary = [[] for _ in range(5)]\n",
    "trn_imgs = [[] for _ in range(2)]\n",
    "with tf.name_scope('train'):\n",
    "    trn_summary[0] = tf.summary.scalar('loss', loss)\n",
    "    trn_summary[1] = tf.summary.histogram(\"p\", tf.nn.softmax(p, axis=0))\n",
    "    for i in range(n_distribution):\n",
    "        trn_summary[2].append(tf.summary.histogram(\"mean/{:d}\".format(i), means[i]))\n",
    "        trn_summary[3].append(tf.summary.histogram(\"cov/{:d}\".format(i), tf.abs(covs[i])))\n",
    "    trn_summary[4] = tf.summary.scalar('gamma', tf.abs(gamma)[0])\n",
    "    image_grid = tf.contrib.gan.eval.image_grid(tf.gather(prep_x(X), np.arange(25)), (5, 5), (28, 28), 1)\n",
    "    trn_imgs[0] = tf.summary.image('input', image_grid, 1)\n",
    "    image_grid = tf.contrib.gan.eval.image_grid(tf.gather(decoder_op, np.arange(25)), (5, 5), (28, 28), 1)\n",
    "    trn_imgs[1] = tf.summary.image('output', image_grid, 1)\n",
    "\n",
    "tst_summary = [[] for _ in range(3)]\n",
    "with tf.name_scope('test'):\n",
    "    tst_summary[0] = tf.summary.scalar('loss', loss)\n",
    "    image_grid = tf.contrib.gan.eval.image_grid(tf.gather(prep_x(X), np.arange(25)), (5, 5), (28, 28), 1)\n",
    "    tst_summary[1] = tf.summary.image('input', image_grid, 1)\n",
    "    image_grid = tf.contrib.gan.eval.image_grid(tf.gather(decoder_op, np.arange(25)), (5, 5), (28, 28), 1)\n",
    "    tst_summary[2] = tf.summary.image('output', image_grid, 1)\n",
    "\n",
    "current_date = datetime.now()\n",
    "current_date = current_date.strftime('%d%b_%H%M%S')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "is_executing": true,
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Loss:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.18669:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.18016:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.19430:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.17887:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.18146:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.18870:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.18734:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.18517:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.17805:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.18342:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.18055:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.17685:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.17333:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.16451:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.16022:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.16056:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.14459:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.13094:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.12013:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.10624:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.09172:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.08270:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.08270:   0%|          | 1/250 [00:02<11:51,  2.86s/it]\n",
      "Loss: 0.08079:   0%|          | 1/250 [00:02<11:51,  2.86s/it]\n",
      "Loss: 0.06939:   0%|          | 1/250 [00:03<11:51,  2.86s/it]\n",
      "Loss: 0.06471:   0%|          | 1/250 [00:03<11:51,  2.86s/it]\n",
      "Loss: 0.06870:   0%|          | 1/250 [00:03<11:51,  2.86s/it]\n",
      "Loss: 0.06761:   0%|          | 1/250 [00:03<11:51,  2.86s/it]\n",
      "Loss: 0.07066:   0%|          | 1/250 [00:03<11:51,  2.86s/it]\n",
      "Loss: 0.06278:   0%|          | 1/250 [00:03<11:51,  2.86s/it]\n",
      "Loss: 0.07413:   0%|          | 1/250 [00:03<11:51,  2.86s/it]\n",
      "Loss: 0.06126:   0%|          | 1/250 [00:03<11:51,  2.86s/it]\n",
      "Loss: 0.07022:   0%|          | 1/250 [00:03<11:51,  2.86s/it]\n",
      "Loss: 0.06999:   0%|          | 1/250 [00:04<11:51,  2.86s/it]\n",
      "Loss: 0.07534:   0%|          | 1/250 [00:04<11:51,  2.86s/it]\n",
      "Loss: 0.06952:   0%|          | 1/250 [00:04<11:51,  2.86s/it]\n",
      "Loss: 0.07423:   0%|          | 1/250 [00:04<11:51,  2.86s/it]\n",
      "Loss: 0.08037:   0%|          | 1/250 [00:04<11:51,  2.86s/it]\n",
      "Loss: 0.07109:   0%|          | 1/250 [00:04<11:51,  2.86s/it]\n",
      "Loss: 0.06359:   0%|          | 1/250 [00:04<11:51,  2.86s/it]\n",
      "Loss: 0.06342:   0%|          | 1/250 [00:04<11:51,  2.86s/it]\n",
      "Loss: 0.05938:   0%|          | 1/250 [00:04<11:51,  2.86s/it]\n",
      "Loss: 0.05633:   0%|          | 1/250 [00:04<11:51,  2.86s/it]\n",
      "Loss: 0.05227:   0%|          | 1/250 [00:05<11:51,  2.86s/it]\n",
      "Loss: 0.05227:   1%|          | 2/250 [00:05<11:01,  2.67s/it]\n",
      "Loss: 0.05692:   1%|          | 2/250 [00:05<11:01,  2.67s/it]\n",
      "Loss: 0.06120:   1%|          | 2/250 [00:05<11:01,  2.67s/it]\n",
      "Loss: 0.05497:   1%|          | 2/250 [00:05<11:01,  2.67s/it]\n",
      "Loss: 0.06079:   1%|          | 2/250 [00:05<11:01,  2.67s/it]\n",
      "Loss: 0.05752:   1%|          | 2/250 [00:05<11:01,  2.67s/it]\n",
      "Loss: 0.05347:   1%|          | 2/250 [00:05<11:01,  2.67s/it]\n",
      "Loss: 0.05358:   1%|          | 2/250 [00:05<11:01,  2.67s/it]\n",
      "Loss: 0.04647:   1%|          | 2/250 [00:05<11:01,  2.67s/it]\n",
      "Loss: 0.05449:   1%|          | 2/250 [00:06<11:01,  2.67s/it]\n",
      "Loss: 0.04500:   1%|          | 2/250 [00:06<11:01,  2.67s/it]\n",
      "Loss: 0.04219:   1%|          | 2/250 [00:06<11:01,  2.67s/it]\n",
      "Loss: 0.03978:   1%|          | 2/250 [00:06<11:01,  2.67s/it]\n",
      "Loss: 0.06134:   1%|          | 2/250 [00:06<11:01,  2.67s/it]\n",
      "Loss: 0.03912:   1%|          | 2/250 [00:06<11:01,  2.67s/it]\n",
      "Loss: 0.03815:   1%|          | 2/250 [00:06<11:01,  2.67s/it]\n",
      "Loss: 0.04390:   1%|          | 2/250 [00:06<11:01,  2.67s/it]\n",
      "Loss: 0.04347:   1%|          | 2/250 [00:06<11:01,  2.67s/it]\n",
      "Loss: 0.03609:   1%|          | 2/250 [00:07<11:01,  2.67s/it]\n",
      "Loss: 0.03961:   1%|          | 2/250 [00:07<11:01,  2.67s/it]\n",
      "Loss: 0.03433:   1%|          | 2/250 [00:07<11:01,  2.67s/it]\n",
      "Loss: 0.03791:   1%|          | 2/250 [00:07<11:01,  2.67s/it]\n",
      "Loss: 0.03791:   1%|          | 3/250 [00:07<10:25,  2.53s/it]\n",
      "Loss: 0.04084:   1%|          | 3/250 [00:07<10:25,  2.53s/it]\n",
      "Loss: 0.03445:   1%|          | 3/250 [00:07<10:25,  2.53s/it]\n",
      "Loss: 0.04516:   1%|          | 3/250 [00:07<10:25,  2.53s/it]\n",
      "Loss: 0.03733:   1%|          | 3/250 [00:07<10:25,  2.53s/it]\n",
      "Loss: 0.03738:   1%|          | 3/250 [00:07<10:25,  2.53s/it]\n",
      "Loss: 0.03606:   1%|          | 3/250 [00:07<10:25,  2.53s/it]\n",
      "Loss: 0.04152:   1%|          | 3/250 [00:08<10:25,  2.53s/it]\n",
      "Loss: 0.03104:   1%|          | 3/250 [00:08<10:25,  2.53s/it]\n",
      "Loss: 0.04155:   1%|          | 3/250 [00:08<10:25,  2.53s/it]\n",
      "Loss: 0.03638:   1%|          | 3/250 [00:08<10:25,  2.53s/it]\n",
      "Loss: 0.03821:   1%|          | 3/250 [00:08<10:25,  2.53s/it]\n",
      "Loss: 0.03153:   1%|          | 3/250 [00:08<10:25,  2.53s/it]\n",
      "Loss: 0.03424:   1%|          | 3/250 [00:08<10:25,  2.53s/it]\n",
      "Loss: 0.03659:   1%|          | 3/250 [00:08<10:25,  2.53s/it]\n",
      "Loss: 0.03131:   1%|          | 3/250 [00:08<10:25,  2.53s/it]\n",
      "Loss: 0.03412:   1%|          | 3/250 [00:09<10:25,  2.53s/it]\n",
      "Loss: 0.03168:   1%|          | 3/250 [00:09<10:25,  2.53s/it]\n",
      "Loss: 0.03145:   1%|          | 3/250 [00:09<10:25,  2.53s/it]\n",
      "Loss: 0.04142:   1%|          | 3/250 [00:09<10:25,  2.53s/it]\n",
      "Loss: 0.03313:   1%|          | 3/250 [00:09<10:25,  2.53s/it]\n",
      "Loss: 0.03362:   1%|          | 3/250 [00:09<10:25,  2.53s/it]\n",
      "Loss: 0.03362:   2%|▏         | 4/250 [00:09<10:01,  2.45s/it]\n",
      "Loss: 0.03314:   2%|▏         | 4/250 [00:09<10:01,  2.45s/it]\n",
      "Loss: 0.03090:   2%|▏         | 4/250 [00:09<10:01,  2.45s/it]\n",
      "Loss: 0.03054:   2%|▏         | 4/250 [00:09<10:01,  2.45s/it]\n",
      "Loss: 0.02788:   2%|▏         | 4/250 [00:09<10:01,  2.45s/it]\n",
      "Loss: 0.02816:   2%|▏         | 4/250 [00:10<10:01,  2.45s/it]\n",
      "Loss: 0.03201:   2%|▏         | 4/250 [00:10<10:01,  2.45s/it]\n",
      "Loss: 0.02901:   2%|▏         | 4/250 [00:10<10:01,  2.45s/it]\n",
      "Loss: 0.03148:   2%|▏         | 4/250 [00:10<10:01,  2.45s/it]\n",
      "Loss: 0.02852:   2%|▏         | 4/250 [00:10<10:01,  2.45s/it]\n",
      "Loss: 0.03161:   2%|▏         | 4/250 [00:10<10:01,  2.45s/it]\n",
      "Loss: 0.03652:   2%|▏         | 4/250 [00:10<10:01,  2.45s/it]\n",
      "Loss: 0.03234:   2%|▏         | 4/250 [00:10<10:01,  2.45s/it]\n",
      "Loss: 0.03605:   2%|▏         | 4/250 [00:10<10:01,  2.45s/it]\n",
      "Loss: 0.03290:   2%|▏         | 4/250 [00:11<10:01,  2.45s/it]\n",
      "Loss: 0.03463:   2%|▏         | 4/250 [00:11<10:01,  2.45s/it]\n",
      "Loss: 0.02911:   2%|▏         | 4/250 [00:11<10:01,  2.45s/it]\n",
      "Loss: 0.02916:   2%|▏         | 4/250 [00:11<10:01,  2.45s/it]\n",
      "Loss: 0.02795:   2%|▏         | 4/250 [00:11<10:01,  2.45s/it]\n",
      "Loss: 0.02913:   2%|▏         | 4/250 [00:11<10:01,  2.45s/it]\n",
      "Loss: 0.02674:   2%|▏         | 4/250 [00:11<10:01,  2.45s/it]\n",
      "Loss: 0.03052:   2%|▏         | 4/250 [00:11<10:01,  2.45s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 55.69it/s]\u001b[A\n",
      "Loss: 0.03052:   2%|▏         | 5/250 [00:11<09:42,  2.38s/it]\n",
      "Loss: 0.03233:   2%|▏         | 5/250 [00:11<09:42,  2.38s/it]\n",
      "Loss: 0.02891:   2%|▏         | 5/250 [00:11<09:42,  2.38s/it]\n",
      "Loss: 0.02989:   2%|▏         | 5/250 [00:12<09:42,  2.38s/it]\n",
      "Loss: 0.02851:   2%|▏         | 5/250 [00:12<09:42,  2.38s/it]\n",
      "Loss: 0.02952:   2%|▏         | 5/250 [00:12<09:42,  2.38s/it]\n",
      "Loss: 0.02986:   2%|▏         | 5/250 [00:12<09:42,  2.38s/it]\n",
      "Loss: 0.02644:   2%|▏         | 5/250 [00:12<09:42,  2.38s/it]\n",
      "Loss: 0.03763:   2%|▏         | 5/250 [00:12<09:42,  2.38s/it]\n",
      "Loss: 0.02348:   2%|▏         | 5/250 [00:12<09:42,  2.38s/it]\n",
      "Loss: 0.03109:   2%|▏         | 5/250 [00:12<09:42,  2.38s/it]\n",
      "Loss: 0.03195:   2%|▏         | 5/250 [00:12<09:42,  2.38s/it]\n",
      "Loss: 0.03201:   2%|▏         | 5/250 [00:13<09:42,  2.38s/it]\n",
      "Loss: 0.03127:   2%|▏         | 5/250 [00:13<09:42,  2.38s/it]\n",
      "Loss: 0.02962:   2%|▏         | 5/250 [00:13<09:42,  2.38s/it]\n",
      "Loss: 0.03276:   2%|▏         | 5/250 [00:13<09:42,  2.38s/it]\n",
      "Loss: 0.02976:   2%|▏         | 5/250 [00:13<09:42,  2.38s/it]\n",
      "Loss: 0.02888:   2%|▏         | 5/250 [00:13<09:42,  2.38s/it]\n",
      "Loss: 0.02479:   2%|▏         | 5/250 [00:13<09:42,  2.38s/it]\n",
      "Loss: 0.03429:   2%|▏         | 5/250 [00:13<09:42,  2.38s/it]\n",
      "Loss: 0.02554:   2%|▏         | 5/250 [00:13<09:42,  2.38s/it]\n",
      "Loss: 0.02848:   2%|▏         | 5/250 [00:13<09:42,  2.38s/it]\n",
      "Loss: 0.02848:   2%|▏         | 6/250 [00:13<09:28,  2.33s/it]\n",
      "Loss: 0.03425:   2%|▏         | 6/250 [00:14<09:28,  2.33s/it]\n",
      "Loss: 0.02752:   2%|▏         | 6/250 [00:14<09:28,  2.33s/it]\n",
      "Loss: 0.02874:   2%|▏         | 6/250 [00:14<09:28,  2.33s/it]\n",
      "Loss: 0.03006:   2%|▏         | 6/250 [00:14<09:28,  2.33s/it]\n",
      "Loss: 0.03462:   2%|▏         | 6/250 [00:14<09:28,  2.33s/it]\n",
      "Loss: 0.02798:   2%|▏         | 6/250 [00:14<09:28,  2.33s/it]\n",
      "Loss: 0.02590:   2%|▏         | 6/250 [00:14<09:28,  2.33s/it]\n",
      "Loss: 0.02807:   2%|▏         | 6/250 [00:14<09:28,  2.33s/it]\n",
      "Loss: 0.02542:   2%|▏         | 6/250 [00:14<09:28,  2.33s/it]\n",
      "Loss: 0.02875:   2%|▏         | 6/250 [00:15<09:28,  2.33s/it]\n",
      "Loss: 0.03047:   2%|▏         | 6/250 [00:15<09:28,  2.33s/it]\n",
      "Loss: 0.02525:   2%|▏         | 6/250 [00:15<09:28,  2.33s/it]\n",
      "Loss: 0.02829:   2%|▏         | 6/250 [00:15<09:28,  2.33s/it]\n",
      "Loss: 0.02546:   2%|▏         | 6/250 [00:15<09:28,  2.33s/it]\n",
      "Loss: 0.02727:   2%|▏         | 6/250 [00:15<09:28,  2.33s/it]\n",
      "Loss: 0.02475:   2%|▏         | 6/250 [00:15<09:28,  2.33s/it]\n",
      "Loss: 0.03071:   2%|▏         | 6/250 [00:15<09:28,  2.33s/it]\n",
      "Loss: 0.02810:   2%|▏         | 6/250 [00:15<09:28,  2.33s/it]\n",
      "Loss: 0.02470:   2%|▏         | 6/250 [00:16<09:28,  2.33s/it]\n",
      "Loss: 0.02411:   2%|▏         | 6/250 [00:16<09:28,  2.33s/it]\n",
      "Loss: 0.02816:   2%|▏         | 6/250 [00:16<09:28,  2.33s/it]\n",
      "Loss: 0.02816:   3%|▎         | 7/250 [00:16<09:17,  2.29s/it]\n",
      "Loss: 0.02816:   3%|▎         | 7/250 [00:16<09:17,  2.29s/it]\n",
      "Loss: 0.02690:   3%|▎         | 7/250 [00:16<09:17,  2.29s/it]\n",
      "Loss: 0.02809:   3%|▎         | 7/250 [00:16<09:17,  2.29s/it]\n",
      "Loss: 0.02827:   3%|▎         | 7/250 [00:16<09:17,  2.29s/it]\n",
      "Loss: 0.02711:   3%|▎         | 7/250 [00:16<09:17,  2.29s/it]\n",
      "Loss: 0.02586:   3%|▎         | 7/250 [00:16<09:17,  2.29s/it]\n",
      "Loss: 0.02779:   3%|▎         | 7/250 [00:16<09:17,  2.29s/it]\n",
      "Loss: 0.02548:   3%|▎         | 7/250 [00:17<09:17,  2.29s/it]\n",
      "Loss: 0.02809:   3%|▎         | 7/250 [00:17<09:17,  2.29s/it]\n",
      "Loss: 0.02594:   3%|▎         | 7/250 [00:17<09:17,  2.29s/it]\n",
      "Loss: 0.02476:   3%|▎         | 7/250 [00:17<09:17,  2.29s/it]\n",
      "Loss: 0.02692:   3%|▎         | 7/250 [00:17<09:17,  2.29s/it]\n",
      "Loss: 0.02544:   3%|▎         | 7/250 [00:17<09:17,  2.29s/it]\n",
      "Loss: 0.02527:   3%|▎         | 7/250 [00:17<09:17,  2.29s/it]\n",
      "Loss: 0.02524:   3%|▎         | 7/250 [00:17<09:17,  2.29s/it]\n",
      "Loss: 0.02768:   3%|▎         | 7/250 [00:17<09:17,  2.29s/it]\n",
      "Loss: 0.02758:   3%|▎         | 7/250 [00:18<09:17,  2.29s/it]\n",
      "Loss: 0.02319:   3%|▎         | 7/250 [00:18<09:17,  2.29s/it]\n",
      "Loss: 0.02835:   3%|▎         | 7/250 [00:18<09:17,  2.29s/it]\n",
      "Loss: 0.02322:   3%|▎         | 7/250 [00:18<09:17,  2.29s/it]\n",
      "Loss: 0.02707:   3%|▎         | 7/250 [00:18<09:17,  2.29s/it]\n",
      "Loss: 0.02707:   3%|▎         | 8/250 [00:18<09:09,  2.27s/it]\n",
      "Loss: 0.02665:   3%|▎         | 8/250 [00:18<09:09,  2.27s/it]\n",
      "Loss: 0.02424:   3%|▎         | 8/250 [00:18<09:09,  2.27s/it]\n",
      "Loss: 0.02778:   3%|▎         | 8/250 [00:18<09:09,  2.27s/it]\n",
      "Loss: 0.02316:   3%|▎         | 8/250 [00:18<09:09,  2.27s/it]\n",
      "Loss: 0.02216:   3%|▎         | 8/250 [00:18<09:09,  2.27s/it]\n",
      "Loss: 0.02657:   3%|▎         | 8/250 [00:19<09:09,  2.27s/it]\n",
      "Loss: 0.02245:   3%|▎         | 8/250 [00:19<09:09,  2.27s/it]\n",
      "Loss: 0.03163:   3%|▎         | 8/250 [00:19<09:09,  2.27s/it]\n",
      "Loss: 0.01949:   3%|▎         | 8/250 [00:19<09:09,  2.27s/it]\n",
      "Loss: 0.02851:   3%|▎         | 8/250 [00:19<09:09,  2.27s/it]\n",
      "Loss: 0.02689:   3%|▎         | 8/250 [00:19<09:09,  2.27s/it]\n",
      "Loss: 0.02852:   3%|▎         | 8/250 [00:19<09:09,  2.27s/it]\n",
      "Loss: 0.02511:   3%|▎         | 8/250 [00:19<09:09,  2.27s/it]\n",
      "Loss: 0.02215:   3%|▎         | 8/250 [00:19<09:09,  2.27s/it]\n",
      "Loss: 0.02306:   3%|▎         | 8/250 [00:20<09:09,  2.27s/it]\n",
      "Loss: 0.02648:   3%|▎         | 8/250 [00:20<09:09,  2.27s/it]\n",
      "Loss: 0.02462:   3%|▎         | 8/250 [00:20<09:09,  2.27s/it]\n",
      "Loss: 0.02439:   3%|▎         | 8/250 [00:20<09:09,  2.27s/it]\n",
      "Loss: 0.02665:   3%|▎         | 8/250 [00:20<09:09,  2.27s/it]\n",
      "Loss: 0.02664:   3%|▎         | 8/250 [00:20<09:09,  2.27s/it]\n",
      "Loss: 0.02498:   3%|▎         | 8/250 [00:20<09:09,  2.27s/it]\n",
      "Loss: 0.02498:   4%|▎         | 9/250 [00:20<09:02,  2.25s/it]\n",
      "Loss: 0.02451:   4%|▎         | 9/250 [00:20<09:02,  2.25s/it]\n",
      "Loss: 0.02323:   4%|▎         | 9/250 [00:20<09:02,  2.25s/it]\n",
      "Loss: 0.02685:   4%|▎         | 9/250 [00:20<09:02,  2.25s/it]\n",
      "Loss: 0.02187:   4%|▎         | 9/250 [00:21<09:02,  2.25s/it]\n",
      "Loss: 0.02216:   4%|▎         | 9/250 [00:21<09:02,  2.25s/it]\n",
      "Loss: 0.02382:   4%|▎         | 9/250 [00:21<09:02,  2.25s/it]\n",
      "Loss: 0.02396:   4%|▎         | 9/250 [00:21<09:02,  2.25s/it]\n",
      "Loss: 0.02170:   4%|▎         | 9/250 [00:21<09:02,  2.25s/it]\n",
      "Loss: 0.02582:   4%|▎         | 9/250 [00:21<09:02,  2.25s/it]\n",
      "Loss: 0.02657:   4%|▎         | 9/250 [00:21<09:02,  2.25s/it]\n",
      "Loss: 0.02397:   4%|▎         | 9/250 [00:21<09:02,  2.25s/it]\n",
      "Loss: 0.02638:   4%|▎         | 9/250 [00:21<09:02,  2.25s/it]\n",
      "Loss: 0.02337:   4%|▎         | 9/250 [00:22<09:02,  2.25s/it]\n",
      "Loss: 0.02261:   4%|▎         | 9/250 [00:22<09:02,  2.25s/it]\n",
      "Loss: 0.02562:   4%|▎         | 9/250 [00:22<09:02,  2.25s/it]\n",
      "Loss: 0.02514:   4%|▎         | 9/250 [00:22<09:02,  2.25s/it]\n",
      "Loss: 0.02507:   4%|▎         | 9/250 [00:22<09:02,  2.25s/it]\n",
      "Loss: 0.02224:   4%|▎         | 9/250 [00:22<09:02,  2.25s/it]\n",
      "Loss: 0.02368:   4%|▎         | 9/250 [00:22<09:02,  2.25s/it]\n",
      "Loss: 0.02338:   4%|▎         | 9/250 [00:22<09:02,  2.25s/it]\n",
      "Loss: 0.02497:   4%|▎         | 9/250 [00:22<09:02,  2.25s/it]\n",
      "Loss: 0.02497:   4%|▍         | 10/250 [00:22<08:56,  2.24s/it]\n",
      "Loss: 0.02424:   4%|▍         | 10/250 [00:22<08:56,  2.24s/it]\n",
      "Loss: 0.02282:   4%|▍         | 10/250 [00:23<08:56,  2.24s/it]\n",
      "Loss: 0.02424:   4%|▍         | 10/250 [00:23<08:56,  2.24s/it]\n",
      "Loss: 0.02112:   4%|▍         | 10/250 [00:23<08:56,  2.24s/it]\n",
      "Loss: 0.02448:   4%|▍         | 10/250 [00:23<08:56,  2.24s/it]\n",
      "Loss: 0.02483:   4%|▍         | 10/250 [00:23<08:56,  2.24s/it]\n",
      "Loss: 0.02265:   4%|▍         | 10/250 [00:23<08:56,  2.24s/it]\n",
      "Loss: 0.02235:   4%|▍         | 10/250 [00:23<08:56,  2.24s/it]\n",
      "Loss: 0.02466:   4%|▍         | 10/250 [00:23<08:56,  2.24s/it]\n",
      "Loss: 0.02370:   4%|▍         | 10/250 [00:23<08:56,  2.24s/it]\n",
      "Loss: 0.02132:   4%|▍         | 10/250 [00:24<08:56,  2.24s/it]\n",
      "Loss: 0.02288:   4%|▍         | 10/250 [00:24<08:56,  2.24s/it]\n",
      "Loss: 0.02346:   4%|▍         | 10/250 [00:24<08:56,  2.24s/it]\n",
      "Loss: 0.02166:   4%|▍         | 10/250 [00:24<08:56,  2.24s/it]\n",
      "Loss: 0.02271:   4%|▍         | 10/250 [00:24<08:56,  2.24s/it]\n",
      "Loss: 0.02362:   4%|▍         | 10/250 [00:24<08:56,  2.24s/it]\n",
      "Loss: 0.02130:   4%|▍         | 10/250 [00:24<08:56,  2.24s/it]\n",
      "Loss: 0.02244:   4%|▍         | 10/250 [00:24<08:56,  2.24s/it]\n",
      "Loss: 0.02326:   4%|▍         | 10/250 [00:24<08:56,  2.24s/it]\n",
      "Loss: 0.02398:   4%|▍         | 10/250 [00:24<08:56,  2.24s/it]\n",
      "Loss: 0.02420:   4%|▍         | 10/250 [00:25<08:56,  2.24s/it]\n",
      "Loss: 0.02420:   4%|▍         | 11/250 [00:25<08:53,  2.23s/it]\n",
      "Loss: 0.02281:   4%|▍         | 11/250 [00:25<08:53,  2.23s/it]\n",
      "Loss: 0.02231:   4%|▍         | 11/250 [00:25<08:53,  2.23s/it]\n",
      "Loss: 0.02386:   4%|▍         | 11/250 [00:25<08:53,  2.23s/it]\n",
      "Loss: 0.02484:   4%|▍         | 11/250 [00:25<08:53,  2.23s/it]\n",
      "Loss: 0.02374:   4%|▍         | 11/250 [00:25<08:53,  2.23s/it]\n",
      "Loss: 0.02316:   4%|▍         | 11/250 [00:25<08:53,  2.23s/it]\n",
      "Loss: 0.02362:   4%|▍         | 11/250 [00:25<08:53,  2.23s/it]\n",
      "Loss: 0.02135:   4%|▍         | 11/250 [00:25<08:53,  2.23s/it]\n",
      "Loss: 0.02363:   4%|▍         | 11/250 [00:26<08:53,  2.23s/it]\n",
      "Loss: 0.02525:   4%|▍         | 11/250 [00:26<08:53,  2.23s/it]\n",
      "Loss: 0.02207:   4%|▍         | 11/250 [00:26<08:53,  2.23s/it]\n",
      "Loss: 0.02240:   4%|▍         | 11/250 [00:26<08:53,  2.23s/it]\n",
      "Loss: 0.02142:   4%|▍         | 11/250 [00:26<08:53,  2.23s/it]\n",
      "Loss: 0.02059:   4%|▍         | 11/250 [00:26<08:53,  2.23s/it]\n",
      "Loss: 0.02185:   4%|▍         | 11/250 [00:26<08:53,  2.23s/it]\n",
      "Loss: 0.01953:   4%|▍         | 11/250 [00:26<08:53,  2.23s/it]\n",
      "Loss: 0.02286:   4%|▍         | 11/250 [00:26<08:53,  2.23s/it]\n",
      "Loss: 0.02180:   4%|▍         | 11/250 [00:26<08:53,  2.23s/it]\n",
      "Loss: 0.02431:   4%|▍         | 11/250 [00:27<08:53,  2.23s/it]\n",
      "Loss: 0.02225:   4%|▍         | 11/250 [00:27<08:53,  2.23s/it]\n",
      "Loss: 0.02286:   4%|▍         | 11/250 [00:27<08:53,  2.23s/it]\n",
      "Loss: 0.02286:   5%|▍         | 12/250 [00:27<08:48,  2.22s/it]\n",
      "Loss: 0.02252:   5%|▍         | 12/250 [00:27<08:48,  2.22s/it]\n",
      "Loss: 0.02199:   5%|▍         | 12/250 [00:27<08:48,  2.22s/it]\n",
      "Loss: 0.02254:   5%|▍         | 12/250 [00:27<08:48,  2.22s/it]\n",
      "Loss: 0.02017:   5%|▍         | 12/250 [00:27<08:48,  2.22s/it]\n",
      "Loss: 0.01823:   5%|▍         | 12/250 [00:27<08:48,  2.22s/it]\n",
      "Loss: 0.02170:   5%|▍         | 12/250 [00:27<08:48,  2.22s/it]\n",
      "Loss: 0.02018:   5%|▍         | 12/250 [00:28<08:48,  2.22s/it]\n",
      "Loss: 0.02271:   5%|▍         | 12/250 [00:28<08:48,  2.22s/it]\n",
      "Loss: 0.02091:   5%|▍         | 12/250 [00:28<08:48,  2.22s/it]\n",
      "Loss: 0.02402:   5%|▍         | 12/250 [00:28<08:48,  2.22s/it]\n",
      "Loss: 0.02217:   5%|▍         | 12/250 [00:28<08:48,  2.22s/it]\n",
      "Loss: 0.02244:   5%|▍         | 12/250 [00:28<08:48,  2.22s/it]\n",
      "Loss: 0.02048:   5%|▍         | 12/250 [00:28<08:48,  2.22s/it]\n",
      "Loss: 0.01945:   5%|▍         | 12/250 [00:28<08:48,  2.22s/it]\n",
      "Loss: 0.02144:   5%|▍         | 12/250 [00:28<08:48,  2.22s/it]\n",
      "Loss: 0.01827:   5%|▍         | 12/250 [00:28<08:48,  2.22s/it]\n",
      "Loss: 0.02212:   5%|▍         | 12/250 [00:29<08:48,  2.22s/it]\n",
      "Loss: 0.02127:   5%|▍         | 12/250 [00:29<08:48,  2.22s/it]\n",
      "Loss: 0.01862:   5%|▍         | 12/250 [00:29<08:48,  2.22s/it]\n",
      "Loss: 0.02042:   5%|▍         | 12/250 [00:29<08:48,  2.22s/it]\n",
      "Loss: 0.02293:   5%|▍         | 12/250 [00:29<08:48,  2.22s/it]\n",
      "Loss: 0.02293:   5%|▌         | 13/250 [00:29<08:45,  2.22s/it]\n",
      "Loss: 0.02124:   5%|▌         | 13/250 [00:29<08:45,  2.22s/it]\n",
      "Loss: 0.02160:   5%|▌         | 13/250 [00:29<08:45,  2.22s/it]\n",
      "Loss: 0.01970:   5%|▌         | 13/250 [00:29<08:45,  2.22s/it]\n",
      "Loss: 0.01995:   5%|▌         | 13/250 [00:29<08:45,  2.22s/it]\n",
      "Loss: 0.01965:   5%|▌         | 13/250 [00:29<08:45,  2.22s/it]\n",
      "Loss: 0.02114:   5%|▌         | 13/250 [00:30<08:45,  2.22s/it]\n",
      "Loss: 0.01879:   5%|▌         | 13/250 [00:30<08:45,  2.22s/it]\n",
      "Loss: 0.02205:   5%|▌         | 13/250 [00:30<08:45,  2.22s/it]\n",
      "Loss: 0.01999:   5%|▌         | 13/250 [00:30<08:45,  2.22s/it]\n",
      "Loss: 0.02354:   5%|▌         | 13/250 [00:30<08:45,  2.22s/it]\n",
      "Loss: 0.02062:   5%|▌         | 13/250 [00:30<08:45,  2.22s/it]\n",
      "Loss: 0.02446:   5%|▌         | 13/250 [00:30<08:45,  2.22s/it]\n",
      "Loss: 0.02081:   5%|▌         | 13/250 [00:30<08:45,  2.22s/it]\n",
      "Loss: 0.02059:   5%|▌         | 13/250 [00:30<08:45,  2.22s/it]\n",
      "Loss: 0.02110:   5%|▌         | 13/250 [00:31<08:45,  2.22s/it]\n",
      "Loss: 0.01894:   5%|▌         | 13/250 [00:31<08:45,  2.22s/it]\n",
      "Loss: 0.02162:   5%|▌         | 13/250 [00:31<08:45,  2.22s/it]\n",
      "Loss: 0.02055:   5%|▌         | 13/250 [00:31<08:45,  2.22s/it]\n",
      "Loss: 0.01822:   5%|▌         | 13/250 [00:31<08:45,  2.22s/it]\n",
      "Loss: 0.01913:   5%|▌         | 13/250 [00:31<08:45,  2.22s/it]\n",
      "Loss: 0.02292:   5%|▌         | 13/250 [00:31<08:45,  2.22s/it]\n",
      "Loss: 0.02292:   6%|▌         | 14/250 [00:31<08:40,  2.21s/it]\n",
      "Loss: 0.02070:   6%|▌         | 14/250 [00:31<08:40,  2.21s/it]\n",
      "Loss: 0.01961:   6%|▌         | 14/250 [00:31<08:40,  2.21s/it]\n",
      "Loss: 0.02242:   6%|▌         | 14/250 [00:31<08:40,  2.21s/it]\n",
      "Loss: 0.02254:   6%|▌         | 14/250 [00:32<08:40,  2.21s/it]\n",
      "Loss: 0.02267:   6%|▌         | 14/250 [00:32<08:40,  2.21s/it]\n",
      "Loss: 0.02061:   6%|▌         | 14/250 [00:32<08:40,  2.21s/it]\n",
      "Loss: 0.02175:   6%|▌         | 14/250 [00:32<08:40,  2.21s/it]\n",
      "Loss: 0.01910:   6%|▌         | 14/250 [00:32<08:40,  2.21s/it]\n",
      "Loss: 0.02244:   6%|▌         | 14/250 [00:32<08:40,  2.21s/it]\n",
      "Loss: 0.02074:   6%|▌         | 14/250 [00:32<08:40,  2.21s/it]\n",
      "Loss: 0.01919:   6%|▌         | 14/250 [00:32<08:40,  2.21s/it]\n",
      "Loss: 0.02188:   6%|▌         | 14/250 [00:32<08:40,  2.21s/it]\n",
      "Loss: 0.02243:   6%|▌         | 14/250 [00:32<08:40,  2.21s/it]\n",
      "Loss: 0.01819:   6%|▌         | 14/250 [00:33<08:40,  2.21s/it]\n",
      "Loss: 0.01998:   6%|▌         | 14/250 [00:33<08:40,  2.21s/it]\n",
      "Loss: 0.02180:   6%|▌         | 14/250 [00:33<08:40,  2.21s/it]\n",
      "Loss: 0.01831:   6%|▌         | 14/250 [00:33<08:40,  2.21s/it]\n",
      "Loss: 0.02329:   6%|▌         | 14/250 [00:33<08:40,  2.21s/it]\n",
      "Loss: 0.02079:   6%|▌         | 14/250 [00:33<08:40,  2.21s/it]\n",
      "Loss: 0.02031:   6%|▌         | 14/250 [00:33<08:40,  2.21s/it]\n",
      "Loss: 0.02220:   6%|▌         | 14/250 [00:33<08:40,  2.21s/it]\n",
      "Loss: 0.02220:   6%|▌         | 15/250 [00:33<08:36,  2.20s/it]\n",
      "Loss: 0.02010:   6%|▌         | 15/250 [00:33<08:36,  2.20s/it]\n",
      "Loss: 0.02050:   6%|▌         | 15/250 [00:34<08:36,  2.20s/it]\n",
      "Loss: 0.02214:   6%|▌         | 15/250 [00:34<08:36,  2.20s/it]\n",
      "Loss: 0.01848:   6%|▌         | 15/250 [00:34<08:36,  2.20s/it]\n",
      "Loss: 0.01837:   6%|▌         | 15/250 [00:34<08:36,  2.20s/it]\n",
      "Loss: 0.02104:   6%|▌         | 15/250 [00:34<08:36,  2.20s/it]\n",
      "Loss: 0.01673:   6%|▌         | 15/250 [00:34<08:36,  2.20s/it]\n",
      "Loss: 0.01992:   6%|▌         | 15/250 [00:34<08:36,  2.20s/it]\n",
      "Loss: 0.02425:   6%|▌         | 15/250 [00:34<08:36,  2.20s/it]\n",
      "Loss: 0.01935:   6%|▌         | 15/250 [00:34<08:36,  2.20s/it]\n",
      "Loss: 0.01908:   6%|▌         | 15/250 [00:34<08:36,  2.20s/it]\n",
      "Loss: 0.01955:   6%|▌         | 15/250 [00:35<08:36,  2.20s/it]\n",
      "Loss: 0.02104:   6%|▌         | 15/250 [00:35<08:36,  2.20s/it]\n",
      "Loss: 0.01778:   6%|▌         | 15/250 [00:35<08:36,  2.20s/it]\n",
      "Loss: 0.01927:   6%|▌         | 15/250 [00:35<08:36,  2.20s/it]\n",
      "Loss: 0.02238:   6%|▌         | 15/250 [00:35<08:36,  2.20s/it]\n",
      "Loss: 0.01828:   6%|▌         | 15/250 [00:35<08:36,  2.20s/it]\n",
      "Loss: 0.02003:   6%|▌         | 15/250 [00:35<08:36,  2.20s/it]\n",
      "Loss: 0.02052:   6%|▌         | 15/250 [00:35<08:36,  2.20s/it]\n",
      "Loss: 0.02211:   6%|▌         | 15/250 [00:35<08:36,  2.20s/it]\n",
      "Loss: 0.02196:   6%|▌         | 15/250 [00:36<08:36,  2.20s/it]\n",
      "Loss: 0.02196:   6%|▋         | 16/250 [00:36<08:36,  2.21s/it]\n",
      "Loss: 0.01996:   6%|▋         | 16/250 [00:36<08:36,  2.21s/it]\n",
      "Loss: 0.01917:   6%|▋         | 16/250 [00:36<08:36,  2.21s/it]\n",
      "Loss: 0.02359:   6%|▋         | 16/250 [00:36<08:36,  2.21s/it]\n",
      "Loss: 0.01796:   6%|▋         | 16/250 [00:36<08:36,  2.21s/it]\n",
      "Loss: 0.01802:   6%|▋         | 16/250 [00:36<08:36,  2.21s/it]\n",
      "Loss: 0.02020:   6%|▋         | 16/250 [00:36<08:36,  2.21s/it]\n",
      "Loss: 0.01826:   6%|▋         | 16/250 [00:36<08:36,  2.21s/it]\n",
      "Loss: 0.02512:   6%|▋         | 16/250 [00:36<08:36,  2.21s/it]\n",
      "Loss: 0.01599:   6%|▋         | 16/250 [00:36<08:36,  2.21s/it]\n",
      "Loss: 0.02319:   6%|▋         | 16/250 [00:37<08:36,  2.21s/it]\n",
      "Loss: 0.02279:   6%|▋         | 16/250 [00:37<08:36,  2.21s/it]\n",
      "Loss: 0.02109:   6%|▋         | 16/250 [00:37<08:36,  2.21s/it]\n",
      "Loss: 0.02235:   6%|▋         | 16/250 [00:37<08:36,  2.21s/it]\n",
      "Loss: 0.02069:   6%|▋         | 16/250 [00:37<08:36,  2.21s/it]\n",
      "Loss: 0.02335:   6%|▋         | 16/250 [00:37<08:36,  2.21s/it]\n",
      "Loss: 0.01885:   6%|▋         | 16/250 [00:37<08:36,  2.21s/it]\n",
      "Loss: 0.01824:   6%|▋         | 16/250 [00:37<08:36,  2.21s/it]\n",
      "Loss: 0.02124:   6%|▋         | 16/250 [00:37<08:36,  2.21s/it]\n",
      "Loss: 0.01987:   6%|▋         | 16/250 [00:38<08:36,  2.21s/it]\n",
      "Loss: 0.02053:   6%|▋         | 16/250 [00:38<08:36,  2.21s/it]\n",
      "Loss: 0.02039:   6%|▋         | 16/250 [00:38<08:36,  2.21s/it]\n",
      "Loss: 0.02039:   7%|▋         | 17/250 [00:38<08:34,  2.21s/it]\n",
      "Loss: 0.01957:   7%|▋         | 17/250 [00:38<08:34,  2.21s/it]\n",
      "Loss: 0.01870:   7%|▋         | 17/250 [00:38<08:34,  2.21s/it]\n",
      "Loss: 0.02210:   7%|▋         | 17/250 [00:38<08:34,  2.21s/it]\n",
      "Loss: 0.01757:   7%|▋         | 17/250 [00:38<08:34,  2.21s/it]\n",
      "Loss: 0.01719:   7%|▋         | 17/250 [00:38<08:34,  2.21s/it]\n",
      "Loss: 0.01967:   7%|▋         | 17/250 [00:38<08:34,  2.21s/it]\n",
      "Loss: 0.01749:   7%|▋         | 17/250 [00:38<08:34,  2.21s/it]\n",
      "Loss: 0.02556:   7%|▋         | 17/250 [00:39<08:34,  2.21s/it]\n",
      "Loss: 0.01613:   7%|▋         | 17/250 [00:39<08:34,  2.21s/it]\n",
      "Loss: 0.02146:   7%|▋         | 17/250 [00:39<08:34,  2.21s/it]\n",
      "Loss: 0.02247:   7%|▋         | 17/250 [00:39<08:34,  2.21s/it]\n",
      "Loss: 0.02048:   7%|▋         | 17/250 [00:39<08:34,  2.21s/it]\n",
      "Loss: 0.02070:   7%|▋         | 17/250 [00:39<08:34,  2.21s/it]\n",
      "Loss: 0.02105:   7%|▋         | 17/250 [00:39<08:34,  2.21s/it]\n",
      "Loss: 0.01998:   7%|▋         | 17/250 [00:39<08:34,  2.21s/it]\n",
      "Loss: 0.01803:   7%|▋         | 17/250 [00:39<08:34,  2.21s/it]\n",
      "Loss: 0.01839:   7%|▋         | 17/250 [00:40<08:34,  2.21s/it]\n",
      "Loss: 0.01835:   7%|▋         | 17/250 [00:40<08:34,  2.21s/it]\n",
      "Loss: 0.01977:   7%|▋         | 17/250 [00:40<08:34,  2.21s/it]\n",
      "Loss: 0.02142:   7%|▋         | 17/250 [00:40<08:34,  2.21s/it]\n",
      "Loss: 0.02022:   7%|▋         | 17/250 [00:40<08:34,  2.21s/it]\n",
      "Loss: 0.02022:   7%|▋         | 18/250 [00:40<08:32,  2.21s/it]\n",
      "Loss: 0.01919:   7%|▋         | 18/250 [00:40<08:32,  2.21s/it]\n",
      "Loss: 0.02031:   7%|▋         | 18/250 [00:40<08:32,  2.21s/it]\n",
      "Loss: 0.02000:   7%|▋         | 18/250 [00:40<08:32,  2.21s/it]\n",
      "Loss: 0.02164:   7%|▋         | 18/250 [00:40<08:32,  2.21s/it]\n",
      "Loss: 0.02079:   7%|▋         | 18/250 [00:40<08:32,  2.21s/it]\n",
      "Loss: 0.01905:   7%|▋         | 18/250 [00:41<08:32,  2.21s/it]\n",
      "Loss: 0.01993:   7%|▋         | 18/250 [00:41<08:32,  2.21s/it]\n",
      "Loss: 0.02031:   7%|▋         | 18/250 [00:41<08:32,  2.21s/it]\n",
      "Loss: 0.02106:   7%|▋         | 18/250 [00:41<08:32,  2.21s/it]\n",
      "Loss: 0.02062:   7%|▋         | 18/250 [00:41<08:32,  2.21s/it]\n",
      "Loss: 0.01939:   7%|▋         | 18/250 [00:41<08:32,  2.21s/it]\n",
      "Loss: 0.02002:   7%|▋         | 18/250 [00:41<08:32,  2.21s/it]\n",
      "Loss: 0.01935:   7%|▋         | 18/250 [00:41<08:32,  2.21s/it]\n",
      "Loss: 0.01720:   7%|▋         | 18/250 [00:41<08:32,  2.21s/it]\n",
      "Loss: 0.01942:   7%|▋         | 18/250 [00:42<08:32,  2.21s/it]\n",
      "Loss: 0.01760:   7%|▋         | 18/250 [00:42<08:32,  2.21s/it]\n",
      "Loss: 0.01867:   7%|▋         | 18/250 [00:42<08:32,  2.21s/it]\n",
      "Loss: 0.01947:   7%|▋         | 18/250 [00:42<08:32,  2.21s/it]\n",
      "Loss: 0.01806:   7%|▋         | 18/250 [00:42<08:32,  2.21s/it]\n",
      "Loss: 0.01817:   7%|▋         | 18/250 [00:42<08:32,  2.21s/it]\n",
      "Loss: 0.01950:   7%|▋         | 18/250 [00:42<08:32,  2.21s/it]\n",
      "Loss: 0.01950:   8%|▊         | 19/250 [00:42<08:27,  2.20s/it]\n",
      "Loss: 0.01941:   8%|▊         | 19/250 [00:42<08:27,  2.20s/it]\n",
      "Loss: 0.01859:   8%|▊         | 19/250 [00:42<08:27,  2.20s/it]\n",
      "Loss: 0.02135:   8%|▊         | 19/250 [00:42<08:27,  2.20s/it]\n",
      "Loss: 0.01686:   8%|▊         | 19/250 [00:43<08:27,  2.20s/it]\n",
      "Loss: 0.01640:   8%|▊         | 19/250 [00:43<08:27,  2.20s/it]\n",
      "Loss: 0.01965:   8%|▊         | 19/250 [00:43<08:27,  2.20s/it]\n",
      "Loss: 0.01577:   8%|▊         | 19/250 [00:43<08:27,  2.20s/it]\n",
      "Loss: 0.02473:   8%|▊         | 19/250 [00:43<08:27,  2.20s/it]\n",
      "Loss: 0.01563:   8%|▊         | 19/250 [00:43<08:27,  2.20s/it]\n",
      "Loss: 0.02159:   8%|▊         | 19/250 [00:43<08:27,  2.20s/it]\n",
      "Loss: 0.02223:   8%|▊         | 19/250 [00:43<08:27,  2.20s/it]\n",
      "Loss: 0.02034:   8%|▊         | 19/250 [00:43<08:27,  2.20s/it]\n",
      "Loss: 0.01950:   8%|▊         | 19/250 [00:43<08:27,  2.20s/it]\n",
      "Loss: 0.02035:   8%|▊         | 19/250 [00:44<08:27,  2.20s/it]\n",
      "Loss: 0.01942:   8%|▊         | 19/250 [00:44<08:27,  2.20s/it]\n",
      "Loss: 0.01795:   8%|▊         | 19/250 [00:44<08:27,  2.20s/it]\n",
      "Loss: 0.01727:   8%|▊         | 19/250 [00:44<08:27,  2.20s/it]\n",
      "Loss: 0.01869:   8%|▊         | 19/250 [00:44<08:27,  2.20s/it]\n",
      "Loss: 0.02021:   8%|▊         | 19/250 [00:44<08:27,  2.20s/it]\n",
      "Loss: 0.02031:   8%|▊         | 19/250 [00:44<08:27,  2.20s/it]\n",
      "Loss: 0.01960:   8%|▊         | 19/250 [00:44<08:27,  2.20s/it]\n",
      "Loss: 0.01960:   8%|▊         | 20/250 [00:44<08:26,  2.20s/it]\n",
      "Loss: 0.01863:   8%|▊         | 20/250 [00:44<08:26,  2.20s/it]\n",
      "Loss: 0.01854:   8%|▊         | 20/250 [00:45<08:26,  2.20s/it]\n",
      "Loss: 0.02054:   8%|▊         | 20/250 [00:45<08:26,  2.20s/it]\n",
      "Loss: 0.01704:   8%|▊         | 20/250 [00:45<08:26,  2.20s/it]\n",
      "Loss: 0.01613:   8%|▊         | 20/250 [00:45<08:26,  2.20s/it]\n",
      "Loss: 0.02038:   8%|▊         | 20/250 [00:45<08:26,  2.20s/it]\n",
      "Loss: 0.01667:   8%|▊         | 20/250 [00:45<08:26,  2.20s/it]\n",
      "Loss: 0.02377:   8%|▊         | 20/250 [00:45<08:26,  2.20s/it]\n",
      "Loss: 0.01586:   8%|▊         | 20/250 [00:45<08:26,  2.20s/it]\n",
      "Loss: 0.02058:   8%|▊         | 20/250 [00:45<08:26,  2.20s/it]\n",
      "Loss: 0.02168:   8%|▊         | 20/250 [00:46<08:26,  2.20s/it]\n",
      "Loss: 0.01991:   8%|▊         | 20/250 [00:46<08:26,  2.20s/it]\n",
      "Loss: 0.01891:   8%|▊         | 20/250 [00:46<08:26,  2.20s/it]\n",
      "Loss: 0.01862:   8%|▊         | 20/250 [00:46<08:26,  2.20s/it]\n",
      "Loss: 0.01895:   8%|▊         | 20/250 [00:46<08:26,  2.20s/it]\n",
      "Loss: 0.01598:   8%|▊         | 20/250 [00:46<08:26,  2.20s/it]\n",
      "Loss: 0.01834:   8%|▊         | 20/250 [00:46<08:26,  2.20s/it]\n",
      "Loss: 0.01893:   8%|▊         | 20/250 [00:46<08:26,  2.20s/it]\n",
      "Loss: 0.01667:   8%|▊         | 20/250 [00:46<08:26,  2.20s/it]\n",
      "Loss: 0.01757:   8%|▊         | 20/250 [00:47<08:26,  2.20s/it]\n",
      "Loss: 0.01909:   8%|▊         | 20/250 [00:47<08:26,  2.20s/it]\n",
      "Loss: 0.01909:   8%|▊         | 21/250 [00:47<08:24,  2.21s/it]\n",
      "Loss: 0.01862:   8%|▊         | 21/250 [00:47<08:24,  2.21s/it]\n",
      "Loss: 0.01978:   8%|▊         | 21/250 [00:47<08:24,  2.21s/it]\n",
      "Loss: 0.02010:   8%|▊         | 21/250 [00:47<08:24,  2.21s/it]\n",
      "Loss: 0.01698:   8%|▊         | 21/250 [00:47<08:24,  2.21s/it]\n",
      "Loss: 0.01587:   8%|▊         | 21/250 [00:47<08:24,  2.21s/it]\n",
      "Loss: 0.02034:   8%|▊         | 21/250 [00:47<08:24,  2.21s/it]\n",
      "Loss: 0.01553:   8%|▊         | 21/250 [00:47<08:24,  2.21s/it]\n",
      "Loss: 0.02447:   8%|▊         | 21/250 [00:47<08:24,  2.21s/it]\n",
      "Loss: 0.01584:   8%|▊         | 21/250 [00:47<08:24,  2.21s/it]\n",
      "Loss: 0.02052:   8%|▊         | 21/250 [00:48<08:24,  2.21s/it]\n",
      "Loss: 0.02173:   8%|▊         | 21/250 [00:48<08:24,  2.21s/it]\n",
      "Loss: 0.01987:   8%|▊         | 21/250 [00:48<08:24,  2.21s/it]\n",
      "Loss: 0.01895:   8%|▊         | 21/250 [00:48<08:24,  2.21s/it]\n",
      "Loss: 0.01929:   8%|▊         | 21/250 [00:48<08:24,  2.21s/it]\n",
      "Loss: 0.01896:   8%|▊         | 21/250 [00:48<08:24,  2.21s/it]\n",
      "Loss: 0.01738:   8%|▊         | 21/250 [00:48<08:24,  2.21s/it]\n",
      "Loss: 0.01753:   8%|▊         | 21/250 [00:48<08:24,  2.21s/it]\n",
      "Loss: 0.01874:   8%|▊         | 21/250 [00:48<08:24,  2.21s/it]\n",
      "Loss: 0.01852:   8%|▊         | 21/250 [00:49<08:24,  2.21s/it]\n",
      "Loss: 0.01996:   8%|▊         | 21/250 [00:49<08:24,  2.21s/it]\n",
      "Loss: 0.01955:   8%|▊         | 21/250 [00:49<08:24,  2.21s/it]\n",
      "Loss: 0.01955:   9%|▉         | 22/250 [00:49<08:20,  2.20s/it]\n",
      "Loss: 0.01808:   9%|▉         | 22/250 [00:49<08:20,  2.20s/it]\n",
      "Loss: 0.01767:   9%|▉         | 22/250 [00:49<08:20,  2.20s/it]\n",
      "Loss: 0.01852:   9%|▉         | 22/250 [00:49<08:20,  2.20s/it]\n",
      "Loss: 0.02109:   9%|▉         | 22/250 [00:49<08:20,  2.20s/it]\n",
      "Loss: 0.02016:   9%|▉         | 22/250 [00:49<08:20,  2.20s/it]\n",
      "Loss: 0.01778:   9%|▉         | 22/250 [00:49<08:20,  2.20s/it]\n",
      "Loss: 0.02000:   9%|▉         | 22/250 [00:49<08:20,  2.20s/it]\n",
      "Loss: 0.01713:   9%|▉         | 22/250 [00:50<08:20,  2.20s/it]\n",
      "Loss: 0.02065:   9%|▉         | 22/250 [00:50<08:20,  2.20s/it]\n",
      "Loss: 0.01781:   9%|▉         | 22/250 [00:50<08:20,  2.20s/it]\n",
      "Loss: 0.01735:   9%|▉         | 22/250 [00:50<08:20,  2.20s/it]\n",
      "Loss: 0.01781:   9%|▉         | 22/250 [00:50<08:20,  2.20s/it]\n",
      "Loss: 0.01915:   9%|▉         | 22/250 [00:50<08:20,  2.20s/it]\n",
      "Loss: 0.01550:   9%|▉         | 22/250 [00:50<08:20,  2.20s/it]\n",
      "Loss: 0.01791:   9%|▉         | 22/250 [00:50<08:20,  2.20s/it]\n",
      "Loss: 0.02108:   9%|▉         | 22/250 [00:50<08:20,  2.20s/it]\n",
      "Loss: 0.01684:   9%|▉         | 22/250 [00:51<08:20,  2.20s/it]\n",
      "Loss: 0.01907:   9%|▉         | 22/250 [00:51<08:20,  2.20s/it]\n",
      "Loss: 0.01926:   9%|▉         | 22/250 [00:51<08:20,  2.20s/it]\n",
      "Loss: 0.01964:   9%|▉         | 22/250 [00:51<08:20,  2.20s/it]\n",
      "Loss: 0.01862:   9%|▉         | 22/250 [00:51<08:20,  2.20s/it]\n",
      "Loss: 0.01862:   9%|▉         | 23/250 [00:51<08:20,  2.21s/it]\n",
      "Loss: 0.01773:   9%|▉         | 23/250 [00:51<08:20,  2.21s/it]\n",
      "Loss: 0.01815:   9%|▉         | 23/250 [00:51<08:20,  2.21s/it]\n",
      "Loss: 0.01997:   9%|▉         | 23/250 [00:51<08:20,  2.21s/it]\n",
      "Loss: 0.01666:   9%|▉         | 23/250 [00:51<08:20,  2.21s/it]\n",
      "Loss: 0.02021:   9%|▉         | 23/250 [00:51<08:20,  2.21s/it]\n",
      "Loss: 0.01729:   9%|▉         | 23/250 [00:52<08:20,  2.21s/it]\n",
      "Loss: 0.01648:   9%|▉         | 23/250 [00:52<08:20,  2.21s/it]\n",
      "Loss: 0.01726:   9%|▉         | 23/250 [00:52<08:20,  2.21s/it]\n",
      "Loss: 0.01676:   9%|▉         | 23/250 [00:52<08:20,  2.21s/it]\n",
      "Loss: 0.02100:   9%|▉         | 23/250 [00:52<08:20,  2.21s/it]\n",
      "Loss: 0.01741:   9%|▉         | 23/250 [00:52<08:20,  2.21s/it]\n",
      "Loss: 0.01922:   9%|▉         | 23/250 [00:52<08:20,  2.21s/it]\n",
      "Loss: 0.01841:   9%|▉         | 23/250 [00:52<08:20,  2.21s/it]\n",
      "Loss: 0.01786:   9%|▉         | 23/250 [00:52<08:20,  2.21s/it]\n",
      "Loss: 0.02038:   9%|▉         | 23/250 [00:53<08:20,  2.21s/it]\n",
      "Loss: 0.01736:   9%|▉         | 23/250 [00:53<08:20,  2.21s/it]\n",
      "Loss: 0.01944:   9%|▉         | 23/250 [00:53<08:20,  2.21s/it]\n",
      "Loss: 0.01704:   9%|▉         | 23/250 [00:53<08:20,  2.21s/it]\n",
      "Loss: 0.01815:   9%|▉         | 23/250 [00:53<08:20,  2.21s/it]\n",
      "Loss: 0.01655:   9%|▉         | 23/250 [00:53<08:20,  2.21s/it]\n",
      "Loss: 0.01883:   9%|▉         | 23/250 [00:53<08:20,  2.21s/it]\n",
      "Loss: 0.01883:  10%|▉         | 24/250 [00:53<08:17,  2.20s/it]\n",
      "Loss: 0.01809:  10%|▉         | 24/250 [00:53<08:17,  2.20s/it]\n",
      "Loss: 0.01788:  10%|▉         | 24/250 [00:53<08:17,  2.20s/it]\n",
      "Loss: 0.01985:  10%|▉         | 24/250 [00:53<08:17,  2.20s/it]\n",
      "Loss: 0.01577:  10%|▉         | 24/250 [00:54<08:17,  2.20s/it]\n",
      "Loss: 0.01708:  10%|▉         | 24/250 [00:54<08:17,  2.20s/it]\n",
      "Loss: 0.01839:  10%|▉         | 24/250 [00:54<08:17,  2.20s/it]\n",
      "Loss: 0.01553:  10%|▉         | 24/250 [00:54<08:17,  2.20s/it]\n",
      "Loss: 0.02277:  10%|▉         | 24/250 [00:54<08:17,  2.20s/it]\n",
      "Loss: 0.01468:  10%|▉         | 24/250 [00:54<08:17,  2.20s/it]\n",
      "Loss: 0.02014:  10%|▉         | 24/250 [00:54<08:17,  2.20s/it]\n",
      "Loss: 0.02176:  10%|▉         | 24/250 [00:54<08:17,  2.20s/it]\n",
      "Loss: 0.01846:  10%|▉         | 24/250 [00:54<08:17,  2.20s/it]\n",
      "Loss: 0.01865:  10%|▉         | 24/250 [00:55<08:17,  2.20s/it]\n",
      "Loss: 0.01860:  10%|▉         | 24/250 [00:55<08:17,  2.20s/it]\n",
      "Loss: 0.01861:  10%|▉         | 24/250 [00:55<08:17,  2.20s/it]\n",
      "Loss: 0.01957:  10%|▉         | 24/250 [00:55<08:17,  2.20s/it]\n",
      "Loss: 0.01619:  10%|▉         | 24/250 [00:55<08:17,  2.20s/it]\n",
      "Loss: 0.01809:  10%|▉         | 24/250 [00:55<08:17,  2.20s/it]\n",
      "Loss: 0.01763:  10%|▉         | 24/250 [00:55<08:17,  2.20s/it]\n",
      "Loss: 0.01734:  10%|▉         | 24/250 [00:55<08:17,  2.20s/it]\n",
      "Loss: 0.01788:  10%|▉         | 24/250 [00:55<08:17,  2.20s/it]\n",
      "Loss: 0.01788:  10%|█         | 25/250 [00:55<08:17,  2.21s/it]\n",
      "Loss: 0.01788:  10%|█         | 25/250 [00:55<08:17,  2.21s/it]\n",
      "Loss: 0.01769:  10%|█         | 25/250 [00:56<08:17,  2.21s/it]\n",
      "Loss: 0.01942:  10%|█         | 25/250 [00:56<08:17,  2.21s/it]\n",
      "Loss: 0.01608:  10%|█         | 25/250 [00:56<08:17,  2.21s/it]\n",
      "Loss: 0.01487:  10%|█         | 25/250 [00:56<08:17,  2.21s/it]\n",
      "Loss: 0.01865:  10%|█         | 25/250 [00:56<08:17,  2.21s/it]\n",
      "Loss: 0.01525:  10%|█         | 25/250 [00:56<08:17,  2.21s/it]\n",
      "Loss: 0.01666:  10%|█         | 25/250 [00:56<08:17,  2.21s/it]\n",
      "Loss: 0.01931:  10%|█         | 25/250 [00:56<08:17,  2.21s/it]\n",
      "Loss: 0.01792:  10%|█         | 25/250 [00:56<08:17,  2.21s/it]\n",
      "Loss: 0.01667:  10%|█         | 25/250 [00:57<08:17,  2.21s/it]\n",
      "Loss: 0.01706:  10%|█         | 25/250 [00:57<08:17,  2.21s/it]\n",
      "Loss: 0.01739:  10%|█         | 25/250 [00:57<08:17,  2.21s/it]\n",
      "Loss: 0.01628:  10%|█         | 25/250 [00:57<08:17,  2.21s/it]\n",
      "Loss: 0.01666:  10%|█         | 25/250 [00:57<08:17,  2.21s/it]\n",
      "Loss: 0.01704:  10%|█         | 25/250 [00:57<08:17,  2.21s/it]\n",
      "Loss: 0.01923:  10%|█         | 25/250 [00:57<08:17,  2.21s/it]\n",
      "Loss: 0.01655:  10%|█         | 25/250 [00:57<08:17,  2.21s/it]\n",
      "Loss: 0.01791:  10%|█         | 25/250 [00:57<08:17,  2.21s/it]\n",
      "Loss: 0.01584:  10%|█         | 25/250 [00:58<08:17,  2.21s/it]\n",
      "Loss: 0.01806:  10%|█         | 25/250 [00:58<08:17,  2.21s/it]\n",
      "Loss: 0.01806:  10%|█         | 26/250 [00:58<08:17,  2.22s/it]\n",
      "Loss: 0.01779:  10%|█         | 26/250 [00:58<08:17,  2.22s/it]\n",
      "Loss: 0.01760:  10%|█         | 26/250 [00:58<08:17,  2.22s/it]\n",
      "Loss: 0.02019:  10%|█         | 26/250 [00:58<08:17,  2.22s/it]\n",
      "Loss: 0.01965:  10%|█         | 26/250 [00:58<08:17,  2.22s/it]\n",
      "Loss: 0.01946:  10%|█         | 26/250 [00:58<08:17,  2.22s/it]\n",
      "Loss: 0.01702:  10%|█         | 26/250 [00:58<08:17,  2.22s/it]\n",
      "Loss: 0.01821:  10%|█         | 26/250 [00:58<08:17,  2.22s/it]\n",
      "Loss: 0.01686:  10%|█         | 26/250 [00:58<08:17,  2.22s/it]\n",
      "Loss: 0.01926:  10%|█         | 26/250 [00:59<08:17,  2.22s/it]\n",
      "Loss: 0.01691:  10%|█         | 26/250 [00:59<08:17,  2.22s/it]\n",
      "Loss: 0.01668:  10%|█         | 26/250 [00:59<08:17,  2.22s/it]\n",
      "Loss: 0.01834:  10%|█         | 26/250 [00:59<08:17,  2.22s/it]\n",
      "Loss: 0.01735:  10%|█         | 26/250 [00:59<08:17,  2.22s/it]\n",
      "Loss: 0.01445:  10%|█         | 26/250 [00:59<08:17,  2.22s/it]\n",
      "Loss: 0.01932:  10%|█         | 26/250 [00:59<08:17,  2.22s/it]\n",
      "Loss: 0.01629:  10%|█         | 26/250 [00:59<08:17,  2.22s/it]\n",
      "Loss: 0.01695:  10%|█         | 26/250 [00:59<08:17,  2.22s/it]\n",
      "Loss: 0.01757:  10%|█         | 26/250 [01:00<08:17,  2.22s/it]\n",
      "Loss: 0.01560:  10%|█         | 26/250 [01:00<08:17,  2.22s/it]\n",
      "Loss: 0.01642:  10%|█         | 26/250 [01:00<08:17,  2.22s/it]\n",
      "Loss: 0.01852:  10%|█         | 26/250 [01:00<08:17,  2.22s/it]\n",
      "Loss: 0.01852:  11%|█         | 27/250 [01:00<08:14,  2.22s/it]\n",
      "Loss: 0.01716:  11%|█         | 27/250 [01:00<08:14,  2.22s/it]\n",
      "Loss: 0.01722:  11%|█         | 27/250 [01:00<08:14,  2.22s/it]\n",
      "Loss: 0.01928:  11%|█         | 27/250 [01:00<08:14,  2.22s/it]\n",
      "Loss: 0.01592:  11%|█         | 27/250 [01:00<08:14,  2.22s/it]\n",
      "Loss: 0.01442:  11%|█         | 27/250 [01:00<08:14,  2.22s/it]\n",
      "Loss: 0.01825:  11%|█         | 27/250 [01:00<08:14,  2.22s/it]\n",
      "Loss: 0.01493:  11%|█         | 27/250 [01:01<08:14,  2.22s/it]\n",
      "Loss: 0.02233:  11%|█         | 27/250 [01:01<08:14,  2.22s/it]\n",
      "Loss: 0.01982:  11%|█         | 27/250 [01:01<08:14,  2.22s/it]\n",
      "Loss: 0.01701:  11%|█         | 27/250 [01:01<08:14,  2.22s/it]\n",
      "Loss: 0.01701:  11%|█         | 27/250 [01:01<08:14,  2.22s/it]\n",
      "Loss: 0.01826:  11%|█         | 27/250 [01:01<08:14,  2.22s/it]\n",
      "Loss: 0.01743:  11%|█         | 27/250 [01:01<08:14,  2.22s/it]\n",
      "Loss: 0.01815:  11%|█         | 27/250 [01:01<08:14,  2.22s/it]\n",
      "Loss: 0.01921:  11%|█         | 27/250 [01:01<08:14,  2.22s/it]\n",
      "Loss: 0.01704:  11%|█         | 27/250 [01:02<08:14,  2.22s/it]\n",
      "Loss: 0.01869:  11%|█         | 27/250 [01:02<08:14,  2.22s/it]\n",
      "Loss: 0.01632:  11%|█         | 27/250 [01:02<08:14,  2.22s/it]\n",
      "Loss: 0.01846:  11%|█         | 27/250 [01:02<08:14,  2.22s/it]\n",
      "Loss: 0.01602:  11%|█         | 27/250 [01:02<08:14,  2.22s/it]\n",
      "Loss: 0.01815:  11%|█         | 27/250 [01:02<08:14,  2.22s/it]\n",
      "Loss: 0.01815:  11%|█         | 28/250 [01:02<08:08,  2.20s/it]\n",
      "Loss: 0.01711:  11%|█         | 28/250 [01:02<08:08,  2.20s/it]\n",
      "Loss: 0.01695:  11%|█         | 28/250 [01:02<08:08,  2.20s/it]\n",
      "Loss: 0.01956:  11%|█         | 28/250 [01:02<08:08,  2.20s/it]\n",
      "Loss: 0.01658:  11%|█         | 28/250 [01:02<08:08,  2.20s/it]\n",
      "Loss: 0.01475:  11%|█         | 28/250 [01:03<08:08,  2.20s/it]\n",
      "Loss: 0.01810:  11%|█         | 28/250 [01:03<08:08,  2.20s/it]\n",
      "Loss: 0.01436:  11%|█         | 28/250 [01:03<08:08,  2.20s/it]\n",
      "Loss: 0.02322:  11%|█         | 28/250 [01:03<08:08,  2.20s/it]\n",
      "Loss: 0.01429:  11%|█         | 28/250 [01:03<08:08,  2.20s/it]\n",
      "Loss: 0.02028:  11%|█         | 28/250 [01:03<08:08,  2.20s/it]\n",
      "Loss: 0.01964:  11%|█         | 28/250 [01:03<08:08,  2.20s/it]\n",
      "Loss: 0.01745:  11%|█         | 28/250 [01:03<08:08,  2.20s/it]\n",
      "Loss: 0.01767:  11%|█         | 28/250 [01:03<08:08,  2.20s/it]\n",
      "Loss: 0.01807:  11%|█         | 28/250 [01:03<08:08,  2.20s/it]\n",
      "Loss: 0.01875:  11%|█         | 28/250 [01:04<08:08,  2.20s/it]\n",
      "Loss: 0.01569:  11%|█         | 28/250 [01:04<08:08,  2.20s/it]\n",
      "Loss: 0.01632:  11%|█         | 28/250 [01:04<08:08,  2.20s/it]\n",
      "Loss: 0.01670:  11%|█         | 28/250 [01:04<08:08,  2.20s/it]\n",
      "Loss: 0.02116:  11%|█         | 28/250 [01:04<08:08,  2.20s/it]\n",
      "Loss: 0.01584:  11%|█         | 28/250 [01:04<08:08,  2.20s/it]\n",
      "Loss: 0.01840:  11%|█         | 28/250 [01:04<08:08,  2.20s/it]\n",
      "Loss: 0.01840:  12%|█▏        | 29/250 [01:04<08:08,  2.21s/it]\n",
      "Loss: 0.01644:  12%|█▏        | 29/250 [01:04<08:08,  2.21s/it]\n",
      "Loss: 0.01716:  12%|█▏        | 29/250 [01:04<08:08,  2.21s/it]\n",
      "Loss: 0.01881:  12%|█▏        | 29/250 [01:05<08:08,  2.21s/it]\n",
      "Loss: 0.01524:  12%|█▏        | 29/250 [01:05<08:08,  2.21s/it]\n",
      "Loss: 0.01428:  12%|█▏        | 29/250 [01:05<08:08,  2.21s/it]\n",
      "Loss: 0.01832:  12%|█▏        | 29/250 [01:05<08:08,  2.21s/it]\n",
      "Loss: 0.01527:  12%|█▏        | 29/250 [01:05<08:08,  2.21s/it]\n",
      "Loss: 0.02179:  12%|█▏        | 29/250 [01:05<08:08,  2.21s/it]\n",
      "Loss: 0.01402:  12%|█▏        | 29/250 [01:05<08:08,  2.21s/it]\n",
      "Loss: 0.01920:  12%|█▏        | 29/250 [01:05<08:08,  2.21s/it]\n",
      "Loss: 0.01913:  12%|█▏        | 29/250 [01:05<08:08,  2.21s/it]\n",
      "Loss: 0.01671:  12%|█▏        | 29/250 [01:05<08:08,  2.21s/it]\n",
      "Loss: 0.01695:  12%|█▏        | 29/250 [01:06<08:08,  2.21s/it]\n",
      "Loss: 0.01452:  12%|█▏        | 29/250 [01:06<08:08,  2.21s/it]\n",
      "Loss: 0.01668:  12%|█▏        | 29/250 [01:06<08:08,  2.21s/it]\n",
      "Loss: 0.01959:  12%|█▏        | 29/250 [01:06<08:08,  2.21s/it]\n",
      "Loss: 0.01578:  12%|█▏        | 29/250 [01:06<08:08,  2.21s/it]\n",
      "Loss: 0.01735:  12%|█▏        | 29/250 [01:06<08:08,  2.21s/it]\n",
      "Loss: 0.01709:  12%|█▏        | 29/250 [01:06<08:08,  2.21s/it]\n",
      "Loss: 0.01580:  12%|█▏        | 29/250 [01:06<08:08,  2.21s/it]\n",
      "Loss: 0.01757:  12%|█▏        | 29/250 [01:06<08:08,  2.21s/it]\n",
      "Loss: 0.01757:  12%|█▏        | 30/250 [01:06<08:06,  2.21s/it]\n",
      "Loss: 0.01709:  12%|█▏        | 30/250 [01:07<08:06,  2.21s/it]\n",
      "Loss: 0.01712:  12%|█▏        | 30/250 [01:07<08:06,  2.21s/it]\n",
      "Loss: 0.01850:  12%|█▏        | 30/250 [01:07<08:06,  2.21s/it]\n",
      "Loss: 0.01511:  12%|█▏        | 30/250 [01:07<08:06,  2.21s/it]\n",
      "Loss: 0.01447:  12%|█▏        | 30/250 [01:07<08:06,  2.21s/it]\n",
      "Loss: 0.01795:  12%|█▏        | 30/250 [01:07<08:06,  2.21s/it]\n",
      "Loss: 0.01433:  12%|█▏        | 30/250 [01:07<08:06,  2.21s/it]\n",
      "Loss: 0.02264:  12%|█▏        | 30/250 [01:07<08:06,  2.21s/it]\n",
      "Loss: 0.01437:  12%|█▏        | 30/250 [01:07<08:06,  2.21s/it]\n",
      "Loss: 0.01955:  12%|█▏        | 30/250 [01:07<08:06,  2.21s/it]\n",
      "Loss: 0.01903:  12%|█▏        | 30/250 [01:08<08:06,  2.21s/it]\n",
      "Loss: 0.01809:  12%|█▏        | 30/250 [01:08<08:06,  2.21s/it]\n",
      "Loss: 0.01752:  12%|█▏        | 30/250 [01:08<08:06,  2.21s/it]\n",
      "Loss: 0.01890:  12%|█▏        | 30/250 [01:08<08:06,  2.21s/it]\n",
      "Loss: 0.01754:  12%|█▏        | 30/250 [01:08<08:06,  2.21s/it]\n",
      "Loss: 0.01521:  12%|█▏        | 30/250 [01:08<08:06,  2.21s/it]\n",
      "Loss: 0.01650:  12%|█▏        | 30/250 [01:08<08:06,  2.21s/it]\n",
      "Loss: 0.01576:  12%|█▏        | 30/250 [01:08<08:06,  2.21s/it]\n",
      "Loss: 0.02088:  12%|█▏        | 30/250 [01:08<08:06,  2.21s/it]\n",
      "Loss: 0.01598:  12%|█▏        | 30/250 [01:09<08:06,  2.21s/it]\n",
      "Loss: 0.01821:  12%|█▏        | 30/250 [01:09<08:06,  2.21s/it]\n",
      "Loss: 0.01821:  12%|█▏        | 31/250 [01:09<08:04,  2.21s/it]\n",
      "Loss: 0.01694:  12%|█▏        | 31/250 [01:09<08:04,  2.21s/it]\n",
      "Loss: 0.01560:  12%|█▏        | 31/250 [01:09<08:04,  2.21s/it]\n",
      "Loss: 0.01765:  12%|█▏        | 31/250 [01:09<08:04,  2.21s/it]\n",
      "Loss: 0.01973:  12%|█▏        | 31/250 [01:09<08:04,  2.21s/it]\n",
      "Loss: 0.01799:  12%|█▏        | 31/250 [01:09<08:04,  2.21s/it]\n",
      "Loss: 0.01608:  12%|█▏        | 31/250 [01:09<08:04,  2.21s/it]\n",
      "Loss: 0.01799:  12%|█▏        | 31/250 [01:09<08:04,  2.21s/it]\n",
      "Loss: 0.01627:  12%|█▏        | 31/250 [01:10<08:04,  2.21s/it]\n",
      "Loss: 0.01852:  12%|█▏        | 31/250 [01:10<08:04,  2.21s/it]\n",
      "Loss: 0.02007:  12%|█▏        | 31/250 [01:10<08:04,  2.21s/it]\n",
      "Loss: 0.01625:  12%|█▏        | 31/250 [01:10<08:04,  2.21s/it]\n",
      "Loss: 0.01798:  12%|█▏        | 31/250 [01:10<08:04,  2.21s/it]\n",
      "Loss: 0.01579:  12%|█▏        | 31/250 [01:10<08:04,  2.21s/it]\n",
      "Loss: 0.01425:  12%|█▏        | 31/250 [01:10<08:04,  2.21s/it]\n",
      "Loss: 0.01730:  12%|█▏        | 31/250 [01:10<08:04,  2.21s/it]\n",
      "Loss: 0.01525:  12%|█▏        | 31/250 [01:10<08:04,  2.21s/it]\n",
      "Loss: 0.01612:  12%|█▏        | 31/250 [01:10<08:04,  2.21s/it]\n",
      "Loss: 0.01584:  12%|█▏        | 31/250 [01:11<08:04,  2.21s/it]\n",
      "Loss: 0.01536:  12%|█▏        | 31/250 [01:11<08:04,  2.21s/it]\n",
      "Loss: 0.01624:  12%|█▏        | 31/250 [01:11<08:04,  2.21s/it]\n",
      "Loss: 0.01718:  12%|█▏        | 31/250 [01:11<08:04,  2.21s/it]\n",
      "Loss: 0.01718:  13%|█▎        | 32/250 [01:11<08:01,  2.21s/it]\n",
      "Loss: 0.01705:  13%|█▎        | 32/250 [01:11<08:01,  2.21s/it]\n",
      "Loss: 0.01644:  13%|█▎        | 32/250 [01:11<08:01,  2.21s/it]\n",
      "Loss: 0.01919:  13%|█▎        | 32/250 [01:11<08:01,  2.21s/it]\n",
      "Loss: 0.01488:  13%|█▎        | 32/250 [01:11<08:01,  2.21s/it]\n",
      "Loss: 0.01377:  13%|█▎        | 32/250 [01:11<08:01,  2.21s/it]\n",
      "Loss: 0.01641:  13%|█▎        | 32/250 [01:11<08:01,  2.21s/it]\n",
      "Loss: 0.01686:  13%|█▎        | 32/250 [01:12<08:01,  2.21s/it]\n",
      "Loss: 0.01604:  13%|█▎        | 32/250 [01:12<08:01,  2.21s/it]\n",
      "Loss: 0.01878:  13%|█▎        | 32/250 [01:12<08:01,  2.21s/it]\n",
      "Loss: 0.01905:  13%|█▎        | 32/250 [01:12<08:01,  2.21s/it]\n",
      "Loss: 0.01757:  13%|█▎        | 32/250 [01:12<08:01,  2.21s/it]\n",
      "Loss: 0.01540:  13%|█▎        | 32/250 [01:12<08:01,  2.21s/it]\n",
      "Loss: 0.01531:  13%|█▎        | 32/250 [01:12<08:01,  2.21s/it]\n",
      "Loss: 0.01443:  13%|█▎        | 32/250 [01:12<08:01,  2.21s/it]\n",
      "Loss: 0.01746:  13%|█▎        | 32/250 [01:12<08:01,  2.21s/it]\n",
      "Loss: 0.01500:  13%|█▎        | 32/250 [01:13<08:01,  2.21s/it]\n",
      "Loss: 0.01669:  13%|█▎        | 32/250 [01:13<08:01,  2.21s/it]\n",
      "Loss: 0.01568:  13%|█▎        | 32/250 [01:13<08:01,  2.21s/it]\n",
      "Loss: 0.01499:  13%|█▎        | 32/250 [01:13<08:01,  2.21s/it]\n",
      "Loss: 0.01586:  13%|█▎        | 32/250 [01:13<08:01,  2.21s/it]\n",
      "Loss: 0.01647:  13%|█▎        | 32/250 [01:13<08:01,  2.21s/it]\n",
      "Loss: 0.01647:  13%|█▎        | 33/250 [01:13<07:57,  2.20s/it]\n",
      "Loss: 0.01631:  13%|█▎        | 33/250 [01:13<07:57,  2.20s/it]\n",
      "Loss: 0.01653:  13%|█▎        | 33/250 [01:13<07:57,  2.20s/it]\n",
      "Loss: 0.01729:  13%|█▎        | 33/250 [01:13<07:57,  2.20s/it]\n",
      "Loss: 0.01892:  13%|█▎        | 33/250 [01:13<07:57,  2.20s/it]\n",
      "Loss: 0.01786:  13%|█▎        | 33/250 [01:14<07:57,  2.20s/it]\n",
      "Loss: 0.01622:  13%|█▎        | 33/250 [01:14<07:57,  2.20s/it]\n",
      "Loss: 0.01398:  13%|█▎        | 33/250 [01:14<07:57,  2.20s/it]\n",
      "Loss: 0.01627:  13%|█▎        | 33/250 [01:14<07:57,  2.20s/it]\n",
      "Loss: 0.01565:  13%|█▎        | 33/250 [01:14<07:57,  2.20s/it]\n",
      "Loss: 0.01986:  13%|█▎        | 33/250 [01:14<07:57,  2.20s/it]\n",
      "Loss: 0.01718:  13%|█▎        | 33/250 [01:14<07:57,  2.20s/it]\n",
      "Loss: 0.01571:  13%|█▎        | 33/250 [01:14<07:57,  2.20s/it]\n",
      "Loss: 0.01650:  13%|█▎        | 33/250 [01:14<07:57,  2.20s/it]\n",
      "Loss: 0.01421:  13%|█▎        | 33/250 [01:15<07:57,  2.20s/it]\n",
      "Loss: 0.01635:  13%|█▎        | 33/250 [01:15<07:57,  2.20s/it]\n",
      "Loss: 0.01437:  13%|█▎        | 33/250 [01:15<07:57,  2.20s/it]\n",
      "Loss: 0.01643:  13%|█▎        | 33/250 [01:15<07:57,  2.20s/it]\n",
      "Loss: 0.01531:  13%|█▎        | 33/250 [01:15<07:57,  2.20s/it]\n",
      "Loss: 0.01777:  13%|█▎        | 33/250 [01:15<07:57,  2.20s/it]\n",
      "Loss: 0.01569:  13%|█▎        | 33/250 [01:15<07:57,  2.20s/it]\n",
      "Loss: 0.01645:  13%|█▎        | 33/250 [01:15<07:57,  2.20s/it]\n",
      "Loss: 0.01645:  14%|█▎        | 34/250 [01:15<07:53,  2.19s/it]\n",
      "Loss: 0.01642:  14%|█▎        | 34/250 [01:15<07:53,  2.19s/it]\n",
      "Loss: 0.01712:  14%|█▎        | 34/250 [01:15<07:53,  2.19s/it]\n",
      "Loss: 0.01593:  14%|█▎        | 34/250 [01:16<07:53,  2.19s/it]\n",
      "Loss: 0.01587:  14%|█▎        | 34/250 [01:16<07:53,  2.19s/it]\n",
      "Loss: 0.01537:  14%|█▎        | 34/250 [01:16<07:53,  2.19s/it]\n",
      "Loss: 0.01379:  14%|█▎        | 34/250 [01:16<07:53,  2.19s/it]\n",
      "Loss: 0.01588:  14%|█▎        | 34/250 [01:16<07:53,  2.19s/it]\n",
      "Loss: 0.01682:  14%|█▎        | 34/250 [01:16<07:53,  2.19s/it]\n",
      "Loss: 0.01561:  14%|█▎        | 34/250 [01:16<07:53,  2.19s/it]\n",
      "Loss: 0.01514:  14%|█▎        | 34/250 [01:16<07:53,  2.19s/it]\n",
      "Loss: 0.01872:  14%|█▎        | 34/250 [01:16<07:53,  2.19s/it]\n",
      "Loss: 0.01638:  14%|█▎        | 34/250 [01:16<07:53,  2.19s/it]\n",
      "Loss: 0.01746:  14%|█▎        | 34/250 [01:17<07:53,  2.19s/it]\n",
      "Loss: 0.01674:  14%|█▎        | 34/250 [01:17<07:53,  2.19s/it]\n",
      "Loss: 0.01626:  14%|█▎        | 34/250 [01:17<07:53,  2.19s/it]\n",
      "Loss: 0.01804:  14%|█▎        | 34/250 [01:17<07:53,  2.19s/it]\n",
      "Loss: 0.01655:  14%|█▎        | 34/250 [01:17<07:53,  2.19s/it]\n",
      "Loss: 0.01760:  14%|█▎        | 34/250 [01:17<07:53,  2.19s/it]\n",
      "Loss: 0.01556:  14%|█▎        | 34/250 [01:17<07:53,  2.19s/it]\n",
      "Loss: 0.01633:  14%|█▎        | 34/250 [01:17<07:53,  2.19s/it]\n",
      "Loss: 0.01559:  14%|█▎        | 34/250 [01:18<07:53,  2.19s/it]\n",
      "Loss: 0.01634:  14%|█▎        | 34/250 [01:18<07:53,  2.19s/it]\n",
      "Loss: 0.01634:  14%|█▍        | 35/250 [01:18<08:05,  2.26s/it]\n",
      "Loss: 0.01581:  14%|█▍        | 35/250 [01:18<08:05,  2.26s/it]\n",
      "Loss: 0.01713:  14%|█▍        | 35/250 [01:18<08:05,  2.26s/it]\n",
      "Loss: 0.01825:  14%|█▍        | 35/250 [01:18<08:05,  2.26s/it]\n",
      "Loss: 0.01438:  14%|█▍        | 35/250 [01:18<08:05,  2.26s/it]\n",
      "Loss: 0.01385:  14%|█▍        | 35/250 [01:18<08:05,  2.26s/it]\n",
      "Loss: 0.01698:  14%|█▍        | 35/250 [01:18<08:05,  2.26s/it]\n",
      "Loss: 0.01359:  14%|█▍        | 35/250 [01:18<08:05,  2.26s/it]\n",
      "Loss: 0.02102:  14%|█▍        | 35/250 [01:18<08:05,  2.26s/it]\n",
      "Loss: 0.01370:  14%|█▍        | 35/250 [01:19<08:05,  2.26s/it]\n",
      "Loss: 0.01865:  14%|█▍        | 35/250 [01:19<08:05,  2.26s/it]\n",
      "Loss: 0.01525:  14%|█▍        | 35/250 [01:19<08:05,  2.26s/it]\n",
      "Loss: 0.01584:  14%|█▍        | 35/250 [01:19<08:05,  2.26s/it]\n",
      "Loss: 0.01596:  14%|█▍        | 35/250 [01:19<08:05,  2.26s/it]\n",
      "Loss: 0.01408:  14%|█▍        | 35/250 [01:19<08:05,  2.26s/it]\n",
      "Loss: 0.01515:  14%|█▍        | 35/250 [01:19<08:05,  2.26s/it]\n",
      "Loss: 0.01834:  14%|█▍        | 35/250 [01:19<08:05,  2.26s/it]\n",
      "Loss: 0.01497:  14%|█▍        | 35/250 [01:19<08:05,  2.26s/it]\n",
      "Loss: 0.01692:  14%|█▍        | 35/250 [01:20<08:05,  2.26s/it]\n",
      "Loss: 0.01673:  14%|█▍        | 35/250 [01:20<08:05,  2.26s/it]\n",
      "Loss: 0.01741:  14%|█▍        | 35/250 [01:20<08:05,  2.26s/it]\n",
      "Loss: 0.01662:  14%|█▍        | 35/250 [01:20<08:05,  2.26s/it]\n",
      "Loss: 0.01662:  14%|█▍        | 36/250 [01:20<08:00,  2.25s/it]\n",
      "Loss: 0.01537:  14%|█▍        | 36/250 [01:20<08:00,  2.25s/it]\n",
      "Loss: 0.01642:  14%|█▍        | 36/250 [01:20<08:00,  2.25s/it]\n",
      "Loss: 0.01617:  14%|█▍        | 36/250 [01:20<08:00,  2.25s/it]\n",
      "Loss: 0.01863:  14%|█▍        | 36/250 [01:20<08:00,  2.25s/it]\n",
      "Loss: 0.01754:  14%|█▍        | 36/250 [01:20<08:00,  2.25s/it]\n",
      "Loss: 0.01594:  14%|█▍        | 36/250 [01:20<08:00,  2.25s/it]\n",
      "Loss: 0.01667:  14%|█▍        | 36/250 [01:21<08:00,  2.25s/it]\n",
      "Loss: 0.01548:  14%|█▍        | 36/250 [01:21<08:00,  2.25s/it]\n",
      "Loss: 0.01840:  14%|█▍        | 36/250 [01:21<08:00,  2.25s/it]\n",
      "Loss: 0.01606:  14%|█▍        | 36/250 [01:21<08:00,  2.25s/it]\n",
      "Loss: 0.01575:  14%|█▍        | 36/250 [01:21<08:00,  2.25s/it]\n",
      "Loss: 0.01578:  14%|█▍        | 36/250 [01:21<08:00,  2.25s/it]\n",
      "Loss: 0.01620:  14%|█▍        | 36/250 [01:21<08:00,  2.25s/it]\n",
      "Loss: 0.01679:  14%|█▍        | 36/250 [01:21<08:00,  2.25s/it]\n",
      "Loss: 0.01756:  14%|█▍        | 36/250 [01:21<08:00,  2.25s/it]\n",
      "Loss: 0.01531:  14%|█▍        | 36/250 [01:22<08:00,  2.25s/it]\n",
      "Loss: 0.01741:  14%|█▍        | 36/250 [01:22<08:00,  2.25s/it]\n",
      "Loss: 0.01525:  14%|█▍        | 36/250 [01:22<08:00,  2.25s/it]\n",
      "Loss: 0.01563:  14%|█▍        | 36/250 [01:22<08:00,  2.25s/it]\n",
      "Loss: 0.01540:  14%|█▍        | 36/250 [01:22<08:00,  2.25s/it]\n",
      "Loss: 0.01604:  14%|█▍        | 36/250 [01:22<08:00,  2.25s/it]\n",
      "Loss: 0.01604:  15%|█▍        | 37/250 [01:22<07:54,  2.23s/it]\n",
      "Loss: 0.01588:  15%|█▍        | 37/250 [01:22<07:54,  2.23s/it]\n",
      "Loss: 0.01618:  15%|█▍        | 37/250 [01:22<07:54,  2.23s/it]\n",
      "Loss: 0.01764:  15%|█▍        | 37/250 [01:22<07:54,  2.23s/it]\n",
      "Loss: 0.01404:  15%|█▍        | 37/250 [01:22<07:54,  2.23s/it]\n",
      "Loss: 0.01337:  15%|█▍        | 37/250 [01:23<07:54,  2.23s/it]\n",
      "Loss: 0.01546:  15%|█▍        | 37/250 [01:23<07:54,  2.23s/it]\n",
      "Loss: 0.01710:  15%|█▍        | 37/250 [01:23<07:54,  2.23s/it]\n",
      "Loss: 0.01542:  15%|█▍        | 37/250 [01:23<07:54,  2.23s/it]\n",
      "Loss: 0.01764:  15%|█▍        | 37/250 [01:23<07:54,  2.23s/it]\n",
      "Loss: 0.01540:  15%|█▍        | 37/250 [01:23<07:54,  2.23s/it]\n",
      "Loss: 0.01561:  15%|█▍        | 37/250 [01:23<07:54,  2.23s/it]\n",
      "Loss: 0.01593:  15%|█▍        | 37/250 [01:23<07:54,  2.23s/it]\n",
      "Loss: 0.01630:  15%|█▍        | 37/250 [01:23<07:54,  2.23s/it]\n",
      "Loss: 0.01563:  15%|█▍        | 37/250 [01:24<07:54,  2.23s/it]\n",
      "Loss: 0.01634:  15%|█▍        | 37/250 [01:24<07:54,  2.23s/it]\n",
      "Loss: 0.01400:  15%|█▍        | 37/250 [01:24<07:54,  2.23s/it]\n",
      "Loss: 0.01571:  15%|█▍        | 37/250 [01:24<07:54,  2.23s/it]\n",
      "Loss: 0.01537:  15%|█▍        | 37/250 [01:24<07:54,  2.23s/it]\n",
      "Loss: 0.01470:  15%|█▍        | 37/250 [01:24<07:54,  2.23s/it]\n",
      "Loss: 0.01517:  15%|█▍        | 37/250 [01:24<07:54,  2.23s/it]\n",
      "Loss: 0.01698:  15%|█▍        | 37/250 [01:24<07:54,  2.23s/it]\n",
      "Loss: 0.01698:  15%|█▌        | 38/250 [01:24<07:51,  2.23s/it]\n",
      "Loss: 0.01522:  15%|█▌        | 38/250 [01:24<07:51,  2.23s/it]\n",
      "Loss: 0.01617:  15%|█▌        | 38/250 [01:24<07:51,  2.23s/it]\n",
      "Loss: 0.01786:  15%|█▌        | 38/250 [01:25<07:51,  2.23s/it]\n",
      "Loss: 0.01425:  15%|█▌        | 38/250 [01:25<07:51,  2.23s/it]\n",
      "Loss: 0.01290:  15%|█▌        | 38/250 [01:25<07:51,  2.23s/it]\n",
      "Loss: 0.01680:  15%|█▌        | 38/250 [01:25<07:51,  2.23s/it]\n",
      "Loss: 0.01359:  15%|█▌        | 38/250 [01:25<07:51,  2.23s/it]\n",
      "Loss: 0.02059:  15%|█▌        | 38/250 [01:25<07:51,  2.23s/it]\n",
      "Loss: 0.01417:  15%|█▌        | 38/250 [01:25<07:51,  2.23s/it]\n",
      "Loss: 0.01783:  15%|█▌        | 38/250 [01:25<07:51,  2.23s/it]\n",
      "Loss: 0.01830:  15%|█▌        | 38/250 [01:25<07:51,  2.23s/it]\n",
      "Loss: 0.01654:  15%|█▌        | 38/250 [01:25<07:51,  2.23s/it]\n",
      "Loss: 0.01637:  15%|█▌        | 38/250 [01:26<07:51,  2.23s/it]\n",
      "Loss: 0.01715:  15%|█▌        | 38/250 [01:26<07:51,  2.23s/it]\n",
      "Loss: 0.01503:  15%|█▌        | 38/250 [01:26<07:51,  2.23s/it]\n",
      "Loss: 0.01808:  15%|█▌        | 38/250 [01:26<07:51,  2.23s/it]\n",
      "Loss: 0.01462:  15%|█▌        | 38/250 [01:26<07:51,  2.23s/it]\n",
      "Loss: 0.01655:  15%|█▌        | 38/250 [01:26<07:51,  2.23s/it]\n",
      "Loss: 0.01686:  15%|█▌        | 38/250 [01:26<07:51,  2.23s/it]\n",
      "Loss: 0.01656:  15%|█▌        | 38/250 [01:26<07:51,  2.23s/it]\n",
      "Loss: 0.01621:  15%|█▌        | 38/250 [01:26<07:51,  2.23s/it]\n",
      "Loss: 0.01621:  16%|█▌        | 39/250 [01:26<07:46,  2.21s/it]\n",
      "Loss: 0.01530:  16%|█▌        | 39/250 [01:27<07:46,  2.21s/it]\n",
      "Loss: 0.01596:  16%|█▌        | 39/250 [01:27<07:46,  2.21s/it]\n",
      "Loss: 0.01779:  16%|█▌        | 39/250 [01:27<07:46,  2.21s/it]\n",
      "Loss: 0.01434:  16%|█▌        | 39/250 [01:27<07:46,  2.21s/it]\n",
      "Loss: 0.01337:  16%|█▌        | 39/250 [01:27<07:46,  2.21s/it]\n",
      "Loss: 0.01568:  16%|█▌        | 39/250 [01:27<07:46,  2.21s/it]\n",
      "Loss: 0.01757:  16%|█▌        | 39/250 [01:27<07:46,  2.21s/it]\n",
      "Loss: 0.01528:  16%|█▌        | 39/250 [01:27<07:46,  2.21s/it]\n",
      "Loss: 0.01756:  16%|█▌        | 39/250 [01:27<07:46,  2.21s/it]\n",
      "Loss: 0.01862:  16%|█▌        | 39/250 [01:27<07:46,  2.21s/it]\n",
      "Loss: 0.01637:  16%|█▌        | 39/250 [01:28<07:46,  2.21s/it]\n",
      "Loss: 0.01682:  16%|█▌        | 39/250 [01:28<07:46,  2.21s/it]\n",
      "Loss: 0.01619:  16%|█▌        | 39/250 [01:28<07:46,  2.21s/it]\n",
      "Loss: 0.01582:  16%|█▌        | 39/250 [01:28<07:46,  2.21s/it]\n",
      "Loss: 0.01638:  16%|█▌        | 39/250 [01:28<07:46,  2.21s/it]\n",
      "Loss: 0.01371:  16%|█▌        | 39/250 [01:28<07:46,  2.21s/it]\n",
      "Loss: 0.01610:  16%|█▌        | 39/250 [01:28<07:46,  2.21s/it]\n",
      "Loss: 0.01565:  16%|█▌        | 39/250 [01:28<07:46,  2.21s/it]\n",
      "Loss: 0.01651:  16%|█▌        | 39/250 [01:28<07:46,  2.21s/it]\n",
      "Loss: 0.01580:  16%|█▌        | 39/250 [01:29<07:46,  2.21s/it]\n",
      "Loss: 0.01685:  16%|█▌        | 39/250 [01:29<07:46,  2.21s/it]\n",
      "Loss: 0.01685:  16%|█▌        | 40/250 [01:29<07:41,  2.20s/it]\n",
      "Loss: 0.01502:  16%|█▌        | 40/250 [01:29<07:41,  2.20s/it]\n",
      "Loss: 0.01571:  16%|█▌        | 40/250 [01:29<07:41,  2.20s/it]\n",
      "Loss: 0.01872:  16%|█▌        | 40/250 [01:29<07:41,  2.20s/it]\n",
      "Loss: 0.01414:  16%|█▌        | 40/250 [01:29<07:41,  2.20s/it]\n",
      "Loss: 0.01283:  16%|█▌        | 40/250 [01:29<07:41,  2.20s/it]\n",
      "Loss: 0.01692:  16%|█▌        | 40/250 [01:29<07:41,  2.20s/it]\n",
      "Loss: 0.01361:  16%|█▌        | 40/250 [01:29<07:41,  2.20s/it]\n",
      "Loss: 0.02066:  16%|█▌        | 40/250 [01:29<07:41,  2.20s/it]\n",
      "Loss: 0.01278:  16%|█▌        | 40/250 [01:30<07:41,  2.20s/it]\n",
      "Loss: 0.01823:  16%|█▌        | 40/250 [01:30<07:41,  2.20s/it]\n",
      "Loss: 0.01759:  16%|█▌        | 40/250 [01:30<07:41,  2.20s/it]\n",
      "Loss: 0.01556:  16%|█▌        | 40/250 [01:30<07:41,  2.20s/it]\n",
      "Loss: 0.01677:  16%|█▌        | 40/250 [01:30<07:41,  2.20s/it]\n",
      "Loss: 0.01706:  16%|█▌        | 40/250 [01:30<07:41,  2.20s/it]\n",
      "Loss: 0.01563:  16%|█▌        | 40/250 [01:30<07:41,  2.20s/it]\n",
      "Loss: 0.01462:  16%|█▌        | 40/250 [01:30<07:41,  2.20s/it]\n",
      "Loss: 0.01513:  16%|█▌        | 40/250 [01:30<07:41,  2.20s/it]\n",
      "Loss: 0.01524:  16%|█▌        | 40/250 [01:31<07:41,  2.20s/it]\n",
      "Loss: 0.02010:  16%|█▌        | 40/250 [01:31<07:41,  2.20s/it]\n",
      "Loss: 0.01502:  16%|█▌        | 40/250 [01:31<07:41,  2.20s/it]\n",
      "Loss: 0.01583:  16%|█▌        | 40/250 [01:31<07:41,  2.20s/it]\n",
      "Loss: 0.01583:  16%|█▋        | 41/250 [01:31<07:42,  2.21s/it]\n",
      "Loss: 0.01584:  16%|█▋        | 41/250 [01:31<07:42,  2.21s/it]\n",
      "Loss: 0.01603:  16%|█▋        | 41/250 [01:31<07:42,  2.21s/it]\n",
      "Loss: 0.01758:  16%|█▋        | 41/250 [01:31<07:42,  2.21s/it]\n",
      "Loss: 0.01411:  16%|█▋        | 41/250 [01:31<07:42,  2.21s/it]\n",
      "Loss: 0.01298:  16%|█▋        | 41/250 [01:31<07:42,  2.21s/it]\n",
      "Loss: 0.01631:  16%|█▋        | 41/250 [01:31<07:42,  2.21s/it]\n",
      "Loss: 0.01338:  16%|█▋        | 41/250 [01:32<07:42,  2.21s/it]\n",
      "Loss: 0.01508:  16%|█▋        | 41/250 [01:32<07:42,  2.21s/it]\n",
      "Loss: 0.01423:  16%|█▋        | 41/250 [01:32<07:42,  2.21s/it]\n",
      "Loss: 0.01798:  16%|█▋        | 41/250 [01:32<07:42,  2.21s/it]\n",
      "Loss: 0.01549:  16%|█▋        | 41/250 [01:32<07:42,  2.21s/it]\n",
      "Loss: 0.01702:  16%|█▋        | 41/250 [01:32<07:42,  2.21s/it]\n",
      "Loss: 0.01580:  16%|█▋        | 41/250 [01:32<07:42,  2.21s/it]\n",
      "Loss: 0.01546:  16%|█▋        | 41/250 [01:32<07:42,  2.21s/it]\n",
      "Loss: 0.01600:  16%|█▋        | 41/250 [01:32<07:42,  2.21s/it]\n",
      "Loss: 0.01393:  16%|█▋        | 41/250 [01:33<07:42,  2.21s/it]\n",
      "Loss: 0.01548:  16%|█▋        | 41/250 [01:33<07:42,  2.21s/it]\n",
      "Loss: 0.01523:  16%|█▋        | 41/250 [01:33<07:42,  2.21s/it]\n",
      "Loss: 0.01470:  16%|█▋        | 41/250 [01:33<07:42,  2.21s/it]\n",
      "Loss: 0.01492:  16%|█▋        | 41/250 [01:33<07:42,  2.21s/it]\n",
      "Loss: 0.01604:  16%|█▋        | 41/250 [01:33<07:42,  2.21s/it]\n",
      "Loss: 0.01604:  17%|█▋        | 42/250 [01:33<07:37,  2.20s/it]\n",
      "Loss: 0.01509:  17%|█▋        | 42/250 [01:33<07:37,  2.20s/it]\n",
      "Loss: 0.01562:  17%|█▋        | 42/250 [01:33<07:37,  2.20s/it]\n",
      "Loss: 0.01637:  17%|█▋        | 42/250 [01:33<07:37,  2.20s/it]\n",
      "Loss: 0.01768:  17%|█▋        | 42/250 [01:33<07:37,  2.20s/it]\n",
      "Loss: 0.01667:  17%|█▋        | 42/250 [01:34<07:37,  2.20s/it]\n",
      "Loss: 0.01557:  17%|█▋        | 42/250 [01:34<07:37,  2.20s/it]\n",
      "Loss: 0.01578:  17%|█▋        | 42/250 [01:34<07:37,  2.20s/it]\n",
      "Loss: 0.01506:  17%|█▋        | 42/250 [01:34<07:37,  2.20s/it]\n",
      "Loss: 0.01722:  17%|█▋        | 42/250 [01:34<07:37,  2.20s/it]\n",
      "Loss: 0.01519:  17%|█▋        | 42/250 [01:34<07:37,  2.20s/it]\n",
      "Loss: 0.01428:  17%|█▋        | 42/250 [01:34<07:37,  2.20s/it]\n",
      "Loss: 0.01576:  17%|█▋        | 42/250 [01:34<07:37,  2.20s/it]\n",
      "Loss: 0.01548:  17%|█▋        | 42/250 [01:34<07:37,  2.20s/it]\n",
      "Loss: 0.01408:  17%|█▋        | 42/250 [01:35<07:37,  2.20s/it]\n",
      "Loss: 0.01485:  17%|█▋        | 42/250 [01:35<07:37,  2.20s/it]\n",
      "Loss: 0.01708:  17%|█▋        | 42/250 [01:35<07:37,  2.20s/it]\n",
      "Loss: 0.01468:  17%|█▋        | 42/250 [01:35<07:37,  2.20s/it]\n",
      "Loss: 0.01496:  17%|█▋        | 42/250 [01:35<07:37,  2.20s/it]\n",
      "Loss: 0.01615:  17%|█▋        | 42/250 [01:35<07:37,  2.20s/it]\n",
      "Loss: 0.01435:  17%|█▋        | 42/250 [01:35<07:37,  2.20s/it]\n",
      "Loss: 0.01595:  17%|█▋        | 42/250 [01:35<07:37,  2.20s/it]\n",
      "Loss: 0.01595:  17%|█▋        | 43/250 [01:35<07:35,  2.20s/it]\n",
      "Loss: 0.01594:  17%|█▋        | 43/250 [01:35<07:35,  2.20s/it]\n",
      "Loss: 0.01523:  17%|█▋        | 43/250 [01:35<07:35,  2.20s/it]\n",
      "Loss: 0.01552:  17%|█▋        | 43/250 [01:36<07:35,  2.20s/it]\n",
      "Loss: 0.01759:  17%|█▋        | 43/250 [01:36<07:35,  2.20s/it]\n",
      "Loss: 0.01648:  17%|█▋        | 43/250 [01:36<07:35,  2.20s/it]\n",
      "Loss: 0.01527:  17%|█▋        | 43/250 [01:36<07:35,  2.20s/it]\n",
      "Loss: 0.01703:  17%|█▋        | 43/250 [01:36<07:35,  2.20s/it]\n",
      "Loss: 0.01492:  17%|█▋        | 43/250 [01:36<07:35,  2.20s/it]\n",
      "Loss: 0.01732:  17%|█▋        | 43/250 [01:36<07:35,  2.20s/it]\n",
      "Loss: 0.01482:  17%|█▋        | 43/250 [01:36<07:35,  2.20s/it]\n",
      "Loss: 0.01453:  17%|█▋        | 43/250 [01:36<07:35,  2.20s/it]\n",
      "Loss: 0.01534:  17%|█▋        | 43/250 [01:37<07:35,  2.20s/it]\n",
      "Loss: 0.01551:  17%|█▋        | 43/250 [01:37<07:35,  2.20s/it]\n",
      "Loss: 0.01553:  17%|█▋        | 43/250 [01:37<07:35,  2.20s/it]\n",
      "Loss: 0.01598:  17%|█▋        | 43/250 [01:37<07:35,  2.20s/it]\n",
      "Loss: 0.01309:  17%|█▋        | 43/250 [01:37<07:35,  2.20s/it]\n",
      "Loss: 0.01611:  17%|█▋        | 43/250 [01:37<07:35,  2.20s/it]\n",
      "Loss: 0.01465:  17%|█▋        | 43/250 [01:37<07:35,  2.20s/it]\n",
      "Loss: 0.01376:  17%|█▋        | 43/250 [01:37<07:35,  2.20s/it]\n",
      "Loss: 0.01490:  17%|█▋        | 43/250 [01:37<07:35,  2.20s/it]\n",
      "Loss: 0.01579:  17%|█▋        | 43/250 [01:37<07:35,  2.20s/it]\n",
      "Loss: 0.01579:  18%|█▊        | 44/250 [01:37<07:34,  2.21s/it]\n",
      "Loss: 0.01490:  18%|█▊        | 44/250 [01:38<07:34,  2.21s/it]\n",
      "Loss: 0.01523:  18%|█▊        | 44/250 [01:38<07:34,  2.21s/it]\n",
      "Loss: 0.01553:  18%|█▊        | 44/250 [01:38<07:34,  2.21s/it]\n",
      "Loss: 0.01740:  18%|█▊        | 44/250 [01:38<07:34,  2.21s/it]\n",
      "Loss: 0.01713:  18%|█▊        | 44/250 [01:38<07:34,  2.21s/it]\n",
      "Loss: 0.01543:  18%|█▊        | 44/250 [01:38<07:34,  2.21s/it]\n",
      "Loss: 0.01405:  18%|█▊        | 44/250 [01:38<07:34,  2.21s/it]\n",
      "Loss: 0.01620:  18%|█▊        | 44/250 [01:38<07:34,  2.21s/it]\n",
      "Loss: 0.01430:  18%|█▊        | 44/250 [01:38<07:34,  2.21s/it]\n",
      "Loss: 0.01486:  18%|█▊        | 44/250 [01:39<07:34,  2.21s/it]\n",
      "Loss: 0.01641:  18%|█▊        | 44/250 [01:39<07:34,  2.21s/it]\n",
      "Loss: 0.01433:  18%|█▊        | 44/250 [01:39<07:34,  2.21s/it]\n",
      "Loss: 0.01404:  18%|█▊        | 44/250 [01:39<07:34,  2.21s/it]\n",
      "Loss: 0.01367:  18%|█▊        | 44/250 [01:39<07:34,  2.21s/it]\n",
      "Loss: 0.01588:  18%|█▊        | 44/250 [01:39<07:34,  2.21s/it]\n",
      "Loss: 0.01374:  18%|█▊        | 44/250 [01:39<07:34,  2.21s/it]\n",
      "Loss: 0.01531:  18%|█▊        | 44/250 [01:39<07:34,  2.21s/it]\n",
      "Loss: 0.01460:  18%|█▊        | 44/250 [01:39<07:34,  2.21s/it]\n",
      "Loss: 0.01405:  18%|█▊        | 44/250 [01:39<07:34,  2.21s/it]\n",
      "Loss: 0.01482:  18%|█▊        | 44/250 [01:40<07:34,  2.21s/it]\n",
      "Loss: 0.01600:  18%|█▊        | 44/250 [01:40<07:34,  2.21s/it]\n",
      "Loss: 0.01600:  18%|█▊        | 45/250 [01:40<07:31,  2.20s/it]\n",
      "Loss: 0.01495:  18%|█▊        | 45/250 [01:40<07:31,  2.20s/it]\n",
      "Loss: 0.01542:  18%|█▊        | 45/250 [01:40<07:31,  2.20s/it]\n",
      "Loss: 0.01769:  18%|█▊        | 45/250 [01:40<07:31,  2.20s/it]\n",
      "Loss: 0.01363:  18%|█▊        | 45/250 [01:40<07:31,  2.20s/it]\n",
      "Loss: 0.01682:  18%|█▊        | 45/250 [01:40<07:31,  2.20s/it]\n",
      "Loss: 0.01520:  18%|█▊        | 45/250 [01:40<07:31,  2.20s/it]\n",
      "Loss: 0.01628:  18%|█▊        | 45/250 [01:40<07:31,  2.20s/it]\n",
      "Loss: 0.01555:  18%|█▊        | 45/250 [01:40<07:31,  2.20s/it]\n",
      "Loss: 0.01702:  18%|█▊        | 45/250 [01:41<07:31,  2.20s/it]\n",
      "Loss: 0.01487:  18%|█▊        | 45/250 [01:41<07:31,  2.20s/it]\n",
      "Loss: 0.01543:  18%|█▊        | 45/250 [01:41<07:31,  2.20s/it]\n",
      "Loss: 0.01631:  18%|█▊        | 45/250 [01:41<07:31,  2.20s/it]\n",
      "Loss: 0.01574:  18%|█▊        | 45/250 [01:41<07:31,  2.20s/it]\n",
      "Loss: 0.01577:  18%|█▊        | 45/250 [01:41<07:31,  2.20s/it]\n",
      "Loss: 0.01664:  18%|█▊        | 45/250 [01:41<07:31,  2.20s/it]\n",
      "Loss: 0.01476:  18%|█▊        | 45/250 [01:41<07:31,  2.20s/it]\n",
      "Loss: 0.01664:  18%|█▊        | 45/250 [01:41<07:31,  2.20s/it]\n",
      "Loss: 0.01477:  18%|█▊        | 45/250 [01:42<07:31,  2.20s/it]\n",
      "Loss: 0.01567:  18%|█▊        | 45/250 [01:42<07:31,  2.20s/it]\n",
      "Loss: 0.01434:  18%|█▊        | 45/250 [01:42<07:31,  2.20s/it]\n",
      "Loss: 0.01545:  18%|█▊        | 45/250 [01:42<07:31,  2.20s/it]\n",
      "Loss: 0.01545:  18%|█▊        | 46/250 [01:42<07:29,  2.20s/it]\n",
      "Loss: 0.01581:  18%|█▊        | 46/250 [01:42<07:29,  2.20s/it]\n",
      "Loss: 0.01530:  18%|█▊        | 46/250 [01:42<07:29,  2.20s/it]\n",
      "Loss: 0.01730:  18%|█▊        | 46/250 [01:42<07:29,  2.20s/it]\n",
      "Loss: 0.01400:  18%|█▊        | 46/250 [01:42<07:29,  2.20s/it]\n",
      "Loss: 0.01283:  18%|█▊        | 46/250 [01:42<07:29,  2.20s/it]\n",
      "Loss: 0.01480:  18%|█▊        | 46/250 [01:42<07:29,  2.20s/it]\n",
      "Loss: 0.01609:  18%|█▊        | 46/250 [01:43<07:29,  2.20s/it]\n",
      "Loss: 0.01429:  18%|█▊        | 46/250 [01:43<07:29,  2.20s/it]\n",
      "Loss: 0.01716:  18%|█▊        | 46/250 [01:43<07:29,  2.20s/it]\n",
      "Loss: 0.01521:  18%|█▊        | 46/250 [01:43<07:29,  2.20s/it]\n",
      "Loss: 0.01437:  18%|█▊        | 46/250 [01:43<07:29,  2.20s/it]\n",
      "Loss: 0.01495:  18%|█▊        | 46/250 [01:43<07:29,  2.20s/it]\n",
      "Loss: 0.01558:  18%|█▊        | 46/250 [01:43<07:29,  2.20s/it]\n",
      "Loss: 0.01307:  18%|█▊        | 46/250 [01:43<07:29,  2.20s/it]\n",
      "Loss: 0.01480:  18%|█▊        | 46/250 [01:43<07:29,  2.20s/it]\n",
      "Loss: 0.01744:  18%|█▊        | 46/250 [01:44<07:29,  2.20s/it]\n",
      "Loss: 0.01405:  18%|█▊        | 46/250 [01:44<07:29,  2.20s/it]\n",
      "Loss: 0.01593:  18%|█▊        | 46/250 [01:44<07:29,  2.20s/it]\n",
      "Loss: 0.01603:  18%|█▊        | 46/250 [01:44<07:29,  2.20s/it]\n",
      "Loss: 0.01570:  18%|█▊        | 46/250 [01:44<07:29,  2.20s/it]\n",
      "Loss: 0.01553:  18%|█▊        | 46/250 [01:44<07:29,  2.20s/it]\n",
      "Loss: 0.01553:  19%|█▉        | 47/250 [01:44<07:28,  2.21s/it]\n",
      "Loss: 0.01523:  19%|█▉        | 47/250 [01:44<07:28,  2.21s/it]\n",
      "Loss: 0.01472:  19%|█▉        | 47/250 [01:44<07:28,  2.21s/it]\n",
      "Loss: 0.01563:  19%|█▉        | 47/250 [01:44<07:28,  2.21s/it]\n",
      "Loss: 0.01708:  19%|█▉        | 47/250 [01:44<07:28,  2.21s/it]\n",
      "Loss: 0.01623:  19%|█▉        | 47/250 [01:45<07:28,  2.21s/it]\n",
      "Loss: 0.01538:  19%|█▉        | 47/250 [01:45<07:28,  2.21s/it]\n",
      "Loss: 0.01606:  19%|█▉        | 47/250 [01:45<07:28,  2.21s/it]\n",
      "Loss: 0.01437:  19%|█▉        | 47/250 [01:45<07:28,  2.21s/it]\n",
      "Loss: 0.01727:  19%|█▉        | 47/250 [01:45<07:28,  2.21s/it]\n",
      "Loss: 0.01524:  19%|█▉        | 47/250 [01:45<07:28,  2.21s/it]\n",
      "Loss: 0.01491:  19%|█▉        | 47/250 [01:45<07:28,  2.21s/it]\n",
      "Loss: 0.01472:  19%|█▉        | 47/250 [01:45<07:28,  2.21s/it]\n",
      "Loss: 0.01568:  19%|█▉        | 47/250 [01:45<07:28,  2.21s/it]\n",
      "Loss: 0.01363:  19%|█▉        | 47/250 [01:46<07:28,  2.21s/it]\n",
      "Loss: 0.01416:  19%|█▉        | 47/250 [01:46<07:28,  2.21s/it]\n",
      "Loss: 0.01440:  19%|█▉        | 47/250 [01:46<07:28,  2.21s/it]\n",
      "Loss: 0.01640:  19%|█▉        | 47/250 [01:46<07:28,  2.21s/it]\n",
      "Loss: 0.01447:  19%|█▉        | 47/250 [01:46<07:28,  2.21s/it]\n",
      "Loss: 0.01569:  19%|█▉        | 47/250 [01:46<07:28,  2.21s/it]\n",
      "Loss: 0.01540:  19%|█▉        | 47/250 [01:46<07:28,  2.21s/it]\n",
      "Loss: 0.01562:  19%|█▉        | 47/250 [01:46<07:28,  2.21s/it]\n",
      "Loss: 0.01562:  19%|█▉        | 48/250 [01:46<07:25,  2.21s/it]\n",
      "Loss: 0.01465:  19%|█▉        | 48/250 [01:46<07:25,  2.21s/it]\n",
      "Loss: 0.01482:  19%|█▉        | 48/250 [01:46<07:25,  2.21s/it]\n",
      "Loss: 0.01673:  19%|█▉        | 48/250 [01:47<07:25,  2.21s/it]\n",
      "Loss: 0.01338:  19%|█▉        | 48/250 [01:47<07:25,  2.21s/it]\n",
      "Loss: 0.01245:  19%|█▉        | 48/250 [01:47<07:25,  2.21s/it]\n",
      "Loss: 0.01588:  19%|█▉        | 48/250 [01:47<07:25,  2.21s/it]\n",
      "Loss: 0.01237:  19%|█▉        | 48/250 [01:47<07:25,  2.21s/it]\n",
      "Loss: 0.01480:  19%|█▉        | 48/250 [01:47<07:25,  2.21s/it]\n",
      "Loss: 0.01682:  19%|█▉        | 48/250 [01:47<07:25,  2.21s/it]\n",
      "Loss: 0.01522:  19%|█▉        | 48/250 [01:47<07:25,  2.21s/it]\n",
      "Loss: 0.01499:  19%|█▉        | 48/250 [01:47<07:25,  2.21s/it]\n",
      "Loss: 0.01489:  19%|█▉        | 48/250 [01:48<07:25,  2.21s/it]\n",
      "Loss: 0.01505:  19%|█▉        | 48/250 [01:48<07:25,  2.21s/it]\n",
      "Loss: 0.01382:  19%|█▉        | 48/250 [01:48<07:25,  2.21s/it]\n",
      "Loss: 0.01401:  19%|█▉        | 48/250 [01:48<07:25,  2.21s/it]\n",
      "Loss: 0.01459:  19%|█▉        | 48/250 [01:48<07:25,  2.21s/it]\n",
      "Loss: 0.01540:  19%|█▉        | 48/250 [01:48<07:25,  2.21s/it]\n",
      "Loss: 0.01402:  19%|█▉        | 48/250 [01:48<07:25,  2.21s/it]\n",
      "Loss: 0.01483:  19%|█▉        | 48/250 [01:48<07:25,  2.21s/it]\n",
      "Loss: 0.01456:  19%|█▉        | 48/250 [01:48<07:25,  2.21s/it]\n",
      "Loss: 0.01611:  19%|█▉        | 48/250 [01:48<07:25,  2.21s/it]\n",
      "Loss: 0.01611:  20%|█▉        | 49/250 [01:48<07:26,  2.22s/it]\n",
      "Loss: 0.01440:  20%|█▉        | 49/250 [01:49<07:26,  2.22s/it]\n",
      "Loss: 0.01493:  20%|█▉        | 49/250 [01:49<07:26,  2.22s/it]\n",
      "Loss: 0.01733:  20%|█▉        | 49/250 [01:49<07:26,  2.22s/it]\n",
      "Loss: 0.01697:  20%|█▉        | 49/250 [01:49<07:26,  2.22s/it]\n",
      "Loss: 0.01643:  20%|█▉        | 49/250 [01:49<07:26,  2.22s/it]\n",
      "Loss: 0.01468:  20%|█▉        | 49/250 [01:49<07:26,  2.22s/it]\n",
      "Loss: 0.01756:  20%|█▉        | 49/250 [01:49<07:26,  2.22s/it]\n",
      "Loss: 0.01452:  20%|█▉        | 49/250 [01:49<07:26,  2.22s/it]\n",
      "Loss: 0.01657:  20%|█▉        | 49/250 [01:49<07:26,  2.22s/it]\n",
      "Loss: 0.01488:  20%|█▉        | 49/250 [01:50<07:26,  2.22s/it]\n",
      "Loss: 0.01405:  20%|█▉        | 49/250 [01:50<07:26,  2.22s/it]\n",
      "Loss: 0.01602:  20%|█▉        | 49/250 [01:50<07:26,  2.22s/it]\n",
      "Loss: 0.01574:  20%|█▉        | 49/250 [01:50<07:26,  2.22s/it]\n",
      "Loss: 0.01558:  20%|█▉        | 49/250 [01:50<07:26,  2.22s/it]\n",
      "Loss: 0.01616:  20%|█▉        | 49/250 [01:50<07:26,  2.22s/it]\n",
      "Loss: 0.01488:  20%|█▉        | 49/250 [01:50<07:26,  2.22s/it]\n",
      "Loss: 0.01650:  20%|█▉        | 49/250 [01:50<07:26,  2.22s/it]\n",
      "Loss: 0.01424:  20%|█▉        | 49/250 [01:50<07:26,  2.22s/it]\n",
      "Loss: 0.01524:  20%|█▉        | 49/250 [01:51<07:26,  2.22s/it]\n",
      "Loss: 0.01408:  20%|█▉        | 49/250 [01:51<07:26,  2.22s/it]\n",
      "Loss: 0.01518:  20%|█▉        | 49/250 [01:51<07:26,  2.22s/it]\n",
      "Loss: 0.01518:  20%|██        | 50/250 [01:51<07:22,  2.21s/it]\n",
      "Loss: 0.01559:  20%|██        | 50/250 [01:51<07:22,  2.21s/it]\n",
      "Loss: 0.01411:  20%|██        | 50/250 [01:51<07:22,  2.21s/it]\n",
      "Loss: 0.01557:  20%|██        | 50/250 [01:51<07:22,  2.21s/it]\n",
      "Loss: 0.01727:  20%|██        | 50/250 [01:51<07:22,  2.21s/it]\n",
      "Loss: 0.01228:  20%|██        | 50/250 [01:51<07:22,  2.21s/it]\n",
      "Loss: 0.01333:  20%|██        | 50/250 [01:51<07:22,  2.21s/it]\n",
      "Loss: 0.01299:  20%|██        | 50/250 [01:51<07:22,  2.21s/it]\n",
      "Loss: 0.01516:  20%|██        | 50/250 [01:52<07:22,  2.21s/it]\n",
      "Loss: 0.01346:  20%|██        | 50/250 [01:52<07:22,  2.21s/it]\n",
      "Loss: 0.01715:  20%|██        | 50/250 [01:52<07:22,  2.21s/it]\n",
      "Loss: 0.01487:  20%|██        | 50/250 [01:52<07:22,  2.21s/it]\n",
      "Loss: 0.01582:  20%|██        | 50/250 [01:52<07:22,  2.21s/it]\n",
      "Loss: 0.01518:  20%|██        | 50/250 [01:52<07:22,  2.21s/it]\n",
      "Loss: 0.01466:  20%|██        | 50/250 [01:52<07:22,  2.21s/it]\n",
      "Loss: 0.01603:  20%|██        | 50/250 [01:52<07:22,  2.21s/it]\n",
      "Loss: 0.01503:  20%|██        | 50/250 [01:52<07:22,  2.21s/it]\n",
      "Loss: 0.01635:  20%|██        | 50/250 [01:52<07:22,  2.21s/it]\n",
      "Loss: 0.01473:  20%|██        | 50/250 [01:53<07:22,  2.21s/it]\n",
      "Loss: 0.01326:  20%|██        | 50/250 [01:53<07:22,  2.21s/it]\n",
      "Loss: 0.01424:  20%|██        | 50/250 [01:53<07:22,  2.21s/it]\n",
      "Loss: 0.01559:  20%|██        | 50/250 [01:53<07:22,  2.21s/it]\n",
      "Loss: 0.01559:  20%|██        | 51/250 [01:53<07:18,  2.20s/it]\n",
      "Loss: 0.01468:  20%|██        | 51/250 [01:53<07:18,  2.20s/it]\n",
      "Loss: 0.01534:  20%|██        | 51/250 [01:53<07:18,  2.20s/it]\n",
      "Loss: 0.01724:  20%|██        | 51/250 [01:53<07:18,  2.20s/it]\n",
      "Loss: 0.01333:  20%|██        | 51/250 [01:53<07:18,  2.20s/it]\n",
      "Loss: 0.01249:  20%|██        | 51/250 [01:53<07:18,  2.20s/it]\n",
      "Loss: 0.01531:  20%|██        | 51/250 [01:53<07:18,  2.20s/it]\n",
      "Loss: 0.01292:  20%|██        | 51/250 [01:54<07:18,  2.20s/it]\n",
      "Loss: 0.01972:  20%|██        | 51/250 [01:54<07:18,  2.20s/it]\n",
      "Loss: 0.01228:  20%|██        | 51/250 [01:54<07:18,  2.20s/it]\n",
      "Loss: 0.01697:  20%|██        | 51/250 [01:54<07:18,  2.20s/it]\n",
      "Loss: 0.01382:  20%|██        | 51/250 [01:54<07:18,  2.20s/it]\n",
      "Loss: 0.01470:  20%|██        | 51/250 [01:54<07:18,  2.20s/it]\n",
      "Loss: 0.01508:  20%|██        | 51/250 [01:54<07:18,  2.20s/it]\n",
      "Loss: 0.01466:  20%|██        | 51/250 [01:54<07:18,  2.20s/it]\n",
      "Loss: 0.01587:  20%|██        | 51/250 [01:54<07:18,  2.20s/it]\n",
      "Loss: 0.01262:  20%|██        | 51/250 [01:55<07:18,  2.20s/it]\n",
      "Loss: 0.01551:  20%|██        | 51/250 [01:55<07:18,  2.20s/it]\n",
      "Loss: 0.01515:  20%|██        | 51/250 [01:55<07:18,  2.20s/it]\n",
      "Loss: 0.01348:  20%|██        | 51/250 [01:55<07:18,  2.20s/it]\n",
      "Loss: 0.01522:  20%|██        | 51/250 [01:55<07:18,  2.20s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 58.98it/s]\u001b[A\n",
      "Loss: 0.01522:  21%|██        | 52/250 [01:55<07:13,  2.19s/it]\n",
      "Loss: 0.01410:  21%|██        | 52/250 [01:55<07:13,  2.19s/it]\n",
      "Loss: 0.01575:  21%|██        | 52/250 [01:55<07:13,  2.19s/it]\n",
      "Loss: 0.01666:  21%|██        | 52/250 [01:55<07:13,  2.19s/it]\n",
      "Loss: 0.01340:  21%|██        | 52/250 [01:55<07:13,  2.19s/it]\n",
      "Loss: 0.01261:  21%|██        | 52/250 [01:56<07:13,  2.19s/it]\n",
      "Loss: 0.01527:  21%|██        | 52/250 [01:56<07:13,  2.19s/it]\n",
      "Loss: 0.01276:  21%|██        | 52/250 [01:56<07:13,  2.19s/it]\n",
      "Loss: 0.01911:  21%|██        | 52/250 [01:56<07:13,  2.19s/it]\n",
      "Loss: 0.01230:  21%|██        | 52/250 [01:56<07:13,  2.19s/it]\n",
      "Loss: 0.01661:  21%|██        | 52/250 [01:56<07:13,  2.19s/it]\n",
      "Loss: 0.01667:  21%|██        | 52/250 [01:56<07:13,  2.19s/it]\n",
      "Loss: 0.01487:  21%|██        | 52/250 [01:56<07:13,  2.19s/it]\n",
      "Loss: 0.01574:  21%|██        | 52/250 [01:56<07:13,  2.19s/it]\n",
      "Loss: 0.01412:  21%|██        | 52/250 [01:57<07:13,  2.19s/it]\n",
      "Loss: 0.01385:  21%|██        | 52/250 [01:57<07:13,  2.19s/it]\n",
      "Loss: 0.01655:  21%|██        | 52/250 [01:57<07:13,  2.19s/it]\n",
      "Loss: 0.01382:  21%|██        | 52/250 [01:57<07:13,  2.19s/it]\n",
      "Loss: 0.01566:  21%|██        | 52/250 [01:57<07:13,  2.19s/it]\n",
      "Loss: 0.01601:  21%|██        | 52/250 [01:57<07:13,  2.19s/it]\n",
      "Loss: 0.01532:  21%|██        | 52/250 [01:57<07:13,  2.19s/it]\n",
      "Loss: 0.01508:  21%|██        | 52/250 [01:57<07:13,  2.19s/it]\n",
      "Loss: 0.01508:  21%|██        | 53/250 [01:57<07:12,  2.20s/it]\n",
      "Loss: 0.01459:  21%|██        | 53/250 [01:57<07:12,  2.20s/it]\n",
      "Loss: 0.01451:  21%|██        | 53/250 [01:57<07:12,  2.20s/it]\n",
      "Loss: 0.01731:  21%|██        | 53/250 [01:58<07:12,  2.20s/it]\n",
      "Loss: 0.01308:  21%|██        | 53/250 [01:58<07:12,  2.20s/it]\n",
      "Loss: 0.01241:  21%|██        | 53/250 [01:58<07:12,  2.20s/it]\n",
      "Loss: 0.01449:  21%|██        | 53/250 [01:58<07:12,  2.20s/it]\n",
      "Loss: 0.01544:  21%|██        | 53/250 [01:58<07:12,  2.20s/it]\n",
      "Loss: 0.01480:  21%|██        | 53/250 [01:58<07:12,  2.20s/it]\n",
      "Loss: 0.01634:  21%|██        | 53/250 [01:58<07:12,  2.20s/it]\n",
      "Loss: 0.01439:  21%|██        | 53/250 [01:58<07:12,  2.20s/it]\n",
      "Loss: 0.01461:  21%|██        | 53/250 [01:58<07:12,  2.20s/it]\n",
      "Loss: 0.01457:  21%|██        | 53/250 [01:59<07:12,  2.20s/it]\n",
      "Loss: 0.01475:  21%|██        | 53/250 [01:59<07:12,  2.20s/it]\n",
      "Loss: 0.01281:  21%|██        | 53/250 [01:59<07:12,  2.20s/it]\n",
      "Loss: 0.01397:  21%|██        | 53/250 [01:59<07:12,  2.20s/it]\n",
      "Loss: 0.01712:  21%|██        | 53/250 [01:59<07:12,  2.20s/it]\n",
      "Loss: 0.01351:  21%|██        | 53/250 [01:59<07:12,  2.20s/it]\n",
      "Loss: 0.01400:  21%|██        | 53/250 [01:59<07:12,  2.20s/it]\n",
      "Loss: 0.01528:  21%|██        | 53/250 [01:59<07:12,  2.20s/it]\n",
      "Loss: 0.01373:  21%|██        | 53/250 [01:59<07:12,  2.20s/it]\n",
      "Loss: 0.01519:  21%|██        | 53/250 [01:59<07:12,  2.20s/it]\n",
      "Loss: 0.01519:  22%|██▏       | 54/250 [01:59<07:11,  2.20s/it]\n",
      "Loss: 0.01458:  22%|██▏       | 54/250 [02:00<07:11,  2.20s/it]\n",
      "Loss: 0.01498:  22%|██▏       | 54/250 [02:00<07:11,  2.20s/it]\n",
      "Loss: 0.01666:  22%|██▏       | 54/250 [02:00<07:11,  2.20s/it]\n",
      "Loss: 0.01309:  22%|██▏       | 54/250 [02:00<07:11,  2.20s/it]\n",
      "Loss: 0.01272:  22%|██▏       | 54/250 [02:00<07:11,  2.20s/it]\n",
      "Loss: 0.01528:  22%|██▏       | 54/250 [02:00<07:11,  2.20s/it]\n",
      "Loss: 0.01237:  22%|██▏       | 54/250 [02:00<07:11,  2.20s/it]\n",
      "Loss: 0.01888:  22%|██▏       | 54/250 [02:00<07:11,  2.20s/it]\n",
      "Loss: 0.01235:  22%|██▏       | 54/250 [02:00<07:11,  2.20s/it]\n",
      "Loss: 0.01677:  22%|██▏       | 54/250 [02:00<07:11,  2.20s/it]\n",
      "Loss: 0.01705:  22%|██▏       | 54/250 [02:01<07:11,  2.20s/it]\n",
      "Loss: 0.01495:  22%|██▏       | 54/250 [02:01<07:11,  2.20s/it]\n",
      "Loss: 0.01564:  22%|██▏       | 54/250 [02:01<07:11,  2.20s/it]\n",
      "Loss: 0.01592:  22%|██▏       | 54/250 [02:01<07:11,  2.20s/it]\n",
      "Loss: 0.01412:  22%|██▏       | 54/250 [02:01<07:11,  2.20s/it]\n",
      "Loss: 0.01429:  22%|██▏       | 54/250 [02:01<07:11,  2.20s/it]\n",
      "Loss: 0.01392:  22%|██▏       | 54/250 [02:01<07:11,  2.20s/it]\n",
      "Loss: 0.01436:  22%|██▏       | 54/250 [02:01<07:11,  2.20s/it]\n",
      "Loss: 0.01630:  22%|██▏       | 54/250 [02:01<07:11,  2.20s/it]\n",
      "Loss: 0.01552:  22%|██▏       | 54/250 [02:02<07:11,  2.20s/it]\n",
      "Loss: 0.01479:  22%|██▏       | 54/250 [02:02<07:11,  2.20s/it]\n",
      "Loss: 0.01479:  22%|██▏       | 55/250 [02:02<07:08,  2.20s/it]\n",
      "Loss: 0.01449:  22%|██▏       | 55/250 [02:02<07:08,  2.20s/it]\n",
      "Loss: 0.01402:  22%|██▏       | 55/250 [02:02<07:08,  2.20s/it]\n",
      "Loss: 0.01497:  22%|██▏       | 55/250 [02:02<07:08,  2.20s/it]\n",
      "Loss: 0.01416:  22%|██▏       | 55/250 [02:02<07:08,  2.20s/it]\n",
      "Loss: 0.01544:  22%|██▏       | 55/250 [02:02<07:08,  2.20s/it]\n",
      "Loss: 0.01544:  22%|██▏       | 55/250 [02:02<07:08,  2.20s/it]\n",
      "Loss: 0.01444:  22%|██▏       | 55/250 [02:02<07:08,  2.20s/it]\n",
      "Loss: 0.01535:  22%|██▏       | 55/250 [02:03<07:08,  2.20s/it]\n",
      "Loss: 0.01651:  22%|██▏       | 55/250 [02:03<07:08,  2.20s/it]\n",
      "Loss: 0.01643:  22%|██▏       | 55/250 [02:03<07:08,  2.20s/it]\n",
      "Loss: 0.01467:  22%|██▏       | 55/250 [02:03<07:08,  2.20s/it]\n",
      "Loss: 0.01382:  22%|██▏       | 55/250 [02:03<07:08,  2.20s/it]\n",
      "Loss: 0.01491:  22%|██▏       | 55/250 [02:03<07:08,  2.20s/it]\n",
      "Loss: 0.01247:  22%|██▏       | 55/250 [02:03<07:08,  2.20s/it]\n",
      "Loss: 0.01355:  22%|██▏       | 55/250 [02:03<07:08,  2.20s/it]\n",
      "Loss: 0.01249:  22%|██▏       | 55/250 [02:03<07:08,  2.20s/it]\n",
      "Loss: 0.01411:  22%|██▏       | 55/250 [02:03<07:08,  2.20s/it]\n",
      "Loss: 0.01510:  22%|██▏       | 55/250 [02:04<07:08,  2.20s/it]\n",
      "Loss: 0.01380:  22%|██▏       | 55/250 [02:04<07:08,  2.20s/it]\n",
      "Loss: 0.01576:  22%|██▏       | 55/250 [02:04<07:08,  2.20s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 57.67it/s]\u001b[A\n",
      "Loss: 0.01576:  22%|██▏       | 56/250 [02:04<07:04,  2.19s/it]\n",
      "Loss: 0.01409:  22%|██▏       | 56/250 [02:04<07:04,  2.19s/it]\n",
      "Loss: 0.01478:  22%|██▏       | 56/250 [02:04<07:04,  2.19s/it]\n",
      "Loss: 0.01631:  22%|██▏       | 56/250 [02:04<07:04,  2.19s/it]\n",
      "Loss: 0.01305:  22%|██▏       | 56/250 [02:04<07:04,  2.19s/it]\n",
      "Loss: 0.01238:  22%|██▏       | 56/250 [02:04<07:04,  2.19s/it]\n",
      "Loss: 0.01454:  22%|██▏       | 56/250 [02:04<07:04,  2.19s/it]\n",
      "Loss: 0.01390:  22%|██▏       | 56/250 [02:05<07:04,  2.19s/it]\n",
      "Loss: 0.01479:  22%|██▏       | 56/250 [02:05<07:04,  2.19s/it]\n",
      "Loss: 0.01328:  22%|██▏       | 56/250 [02:05<07:04,  2.19s/it]\n",
      "Loss: 0.01740:  22%|██▏       | 56/250 [02:05<07:04,  2.19s/it]\n",
      "Loss: 0.01543:  22%|██▏       | 56/250 [02:05<07:04,  2.19s/it]\n",
      "Loss: 0.01529:  22%|██▏       | 56/250 [02:05<07:04,  2.19s/it]\n",
      "Loss: 0.01489:  22%|██▏       | 56/250 [02:05<07:04,  2.19s/it]\n",
      "Loss: 0.01492:  22%|██▏       | 56/250 [02:05<07:04,  2.19s/it]\n",
      "Loss: 0.01563:  22%|██▏       | 56/250 [02:05<07:04,  2.19s/it]\n",
      "Loss: 0.01401:  22%|██▏       | 56/250 [02:06<07:04,  2.19s/it]\n",
      "Loss: 0.01567:  22%|██▏       | 56/250 [02:06<07:04,  2.19s/it]\n",
      "Loss: 0.01432:  22%|██▏       | 56/250 [02:06<07:04,  2.19s/it]\n",
      "Loss: 0.01510:  22%|██▏       | 56/250 [02:06<07:04,  2.19s/it]\n",
      "Loss: 0.01368:  22%|██▏       | 56/250 [02:06<07:04,  2.19s/it]\n",
      "Loss: 0.01487:  22%|██▏       | 56/250 [02:06<07:04,  2.19s/it]\n",
      "Loss: 0.01487:  23%|██▎       | 57/250 [02:06<07:04,  2.20s/it]\n",
      "Loss: 0.01396:  23%|██▎       | 57/250 [02:06<07:04,  2.20s/it]\n",
      "Loss: 0.01494:  23%|██▎       | 57/250 [02:06<07:04,  2.20s/it]\n",
      "Loss: 0.01634:  23%|██▎       | 57/250 [02:06<07:04,  2.20s/it]\n",
      "Loss: 0.01269:  23%|██▎       | 57/250 [02:06<07:04,  2.20s/it]\n",
      "Loss: 0.01220:  23%|██▎       | 57/250 [02:07<07:04,  2.20s/it]\n",
      "Loss: 0.01527:  23%|██▎       | 57/250 [02:07<07:04,  2.20s/it]\n",
      "Loss: 0.01194:  23%|██▎       | 57/250 [02:07<07:04,  2.20s/it]\n",
      "Loss: 0.01870:  23%|██▎       | 57/250 [02:07<07:04,  2.20s/it]\n",
      "Loss: 0.01202:  23%|██▎       | 57/250 [02:07<07:04,  2.20s/it]\n",
      "Loss: 0.01672:  23%|██▎       | 57/250 [02:07<07:04,  2.20s/it]\n",
      "Loss: 0.01699:  23%|██▎       | 57/250 [02:07<07:04,  2.20s/it]\n",
      "Loss: 0.01484:  23%|██▎       | 57/250 [02:07<07:04,  2.20s/it]\n",
      "Loss: 0.01507:  23%|██▎       | 57/250 [02:07<07:04,  2.20s/it]\n",
      "Loss: 0.01598:  23%|██▎       | 57/250 [02:07<07:04,  2.20s/it]\n",
      "Loss: 0.01378:  23%|██▎       | 57/250 [02:08<07:04,  2.20s/it]\n",
      "Loss: 0.01272:  23%|██▎       | 57/250 [02:08<07:04,  2.20s/it]\n",
      "Loss: 0.01426:  23%|██▎       | 57/250 [02:08<07:04,  2.20s/it]\n",
      "Loss: 0.01425:  23%|██▎       | 57/250 [02:08<07:04,  2.20s/it]\n",
      "Loss: 0.01861:  23%|██▎       | 57/250 [02:08<07:04,  2.20s/it]\n",
      "Loss: 0.01374:  23%|██▎       | 57/250 [02:08<07:04,  2.20s/it]\n",
      "Loss: 0.01515:  23%|██▎       | 57/250 [02:08<07:04,  2.20s/it]\n",
      "Loss: 0.01515:  23%|██▎       | 58/250 [02:08<07:01,  2.20s/it]\n",
      "Loss: 0.01383:  23%|██▎       | 58/250 [02:08<07:01,  2.20s/it]\n",
      "Loss: 0.01428:  23%|██▎       | 58/250 [02:08<07:01,  2.20s/it]\n",
      "Loss: 0.01599:  23%|██▎       | 58/250 [02:09<07:01,  2.20s/it]\n",
      "Loss: 0.01329:  23%|██▎       | 58/250 [02:09<07:01,  2.20s/it]\n",
      "Loss: 0.01203:  23%|██▎       | 58/250 [02:09<07:01,  2.20s/it]\n",
      "Loss: 0.01419:  23%|██▎       | 58/250 [02:09<07:01,  2.20s/it]\n",
      "Loss: 0.01536:  23%|██▎       | 58/250 [02:09<07:01,  2.20s/it]\n",
      "Loss: 0.01466:  23%|██▎       | 58/250 [02:09<07:01,  2.20s/it]\n",
      "Loss: 0.01299:  23%|██▎       | 58/250 [02:09<07:01,  2.20s/it]\n",
      "Loss: 0.01727:  23%|██▎       | 58/250 [02:09<07:01,  2.20s/it]\n",
      "Loss: 0.01443:  23%|██▎       | 58/250 [02:09<07:01,  2.20s/it]\n",
      "Loss: 0.01560:  23%|██▎       | 58/250 [02:09<07:01,  2.20s/it]\n",
      "Loss: 0.01474:  23%|██▎       | 58/250 [02:10<07:01,  2.20s/it]\n",
      "Loss: 0.01449:  23%|██▎       | 58/250 [02:10<07:01,  2.20s/it]\n",
      "Loss: 0.01608:  23%|██▎       | 58/250 [02:10<07:01,  2.20s/it]\n",
      "Loss: 0.01411:  23%|██▎       | 58/250 [02:10<07:01,  2.20s/it]\n",
      "Loss: 0.01591:  23%|██▎       | 58/250 [02:10<07:01,  2.20s/it]\n",
      "Loss: 0.01383:  23%|██▎       | 58/250 [02:10<07:01,  2.20s/it]\n",
      "Loss: 0.01529:  23%|██▎       | 58/250 [02:10<07:01,  2.20s/it]\n",
      "Loss: 0.01375:  23%|██▎       | 58/250 [02:10<07:01,  2.20s/it]\n",
      "Loss: 0.01497:  23%|██▎       | 58/250 [02:10<07:01,  2.20s/it]\n",
      "Loss: 0.01497:  24%|██▎       | 59/250 [02:10<06:57,  2.19s/it]\n",
      "Loss: 0.01414:  24%|██▎       | 59/250 [02:10<06:57,  2.19s/it]\n",
      "Loss: 0.01487:  24%|██▎       | 59/250 [02:11<06:57,  2.19s/it]\n",
      "Loss: 0.01644:  24%|██▎       | 59/250 [02:11<06:57,  2.19s/it]\n",
      "Loss: 0.01637:  24%|██▎       | 59/250 [02:11<06:57,  2.19s/it]\n",
      "Loss: 0.01611:  24%|██▎       | 59/250 [02:11<06:57,  2.19s/it]\n",
      "Loss: 0.01454:  24%|██▎       | 59/250 [02:11<06:57,  2.19s/it]\n",
      "Loss: 0.01542:  24%|██▎       | 59/250 [02:11<06:57,  2.19s/it]\n",
      "Loss: 0.01452:  24%|██▎       | 59/250 [02:11<06:57,  2.19s/it]\n",
      "Loss: 0.01636:  24%|██▎       | 59/250 [02:11<06:57,  2.19s/it]\n",
      "Loss: 0.01414:  24%|██▎       | 59/250 [02:11<06:57,  2.19s/it]\n",
      "Loss: 0.01397:  24%|██▎       | 59/250 [02:12<06:57,  2.19s/it]\n",
      "Loss: 0.01387:  24%|██▎       | 59/250 [02:12<06:57,  2.19s/it]\n",
      "Loss: 0.01510:  24%|██▎       | 59/250 [02:12<06:57,  2.19s/it]\n",
      "Loss: 0.01454:  24%|██▎       | 59/250 [02:12<06:57,  2.19s/it]\n",
      "Loss: 0.01571:  24%|██▎       | 59/250 [02:12<06:57,  2.19s/it]\n",
      "Loss: 0.01414:  24%|██▎       | 59/250 [02:12<06:57,  2.19s/it]\n",
      "Loss: 0.01575:  24%|██▎       | 59/250 [02:12<06:57,  2.19s/it]\n",
      "Loss: 0.01399:  24%|██▎       | 59/250 [02:12<06:57,  2.19s/it]\n",
      "Loss: 0.01436:  24%|██▎       | 59/250 [02:12<06:57,  2.19s/it]\n",
      "Loss: 0.01406:  24%|██▎       | 59/250 [02:13<06:57,  2.19s/it]\n",
      "Loss: 0.01472:  24%|██▎       | 59/250 [02:13<06:57,  2.19s/it]\n",
      "Loss: 0.01472:  24%|██▍       | 60/250 [02:13<06:55,  2.19s/it]\n",
      "Loss: 0.01381:  24%|██▍       | 60/250 [02:13<06:55,  2.19s/it]\n",
      "Loss: 0.01445:  24%|██▍       | 60/250 [02:13<06:55,  2.19s/it]\n",
      "Loss: 0.01682:  24%|██▍       | 60/250 [02:13<06:55,  2.19s/it]\n",
      "Loss: 0.01312:  24%|██▍       | 60/250 [02:13<06:55,  2.19s/it]\n",
      "Loss: 0.01212:  24%|██▍       | 60/250 [02:13<06:55,  2.19s/it]\n",
      "Loss: 0.01484:  24%|██▍       | 60/250 [02:13<06:55,  2.19s/it]\n",
      "Loss: 0.01216:  24%|██▍       | 60/250 [02:13<06:55,  2.19s/it]\n",
      "Loss: 0.01901:  24%|██▍       | 60/250 [02:13<06:55,  2.19s/it]\n",
      "Loss: 0.01618:  24%|██▍       | 60/250 [02:14<06:55,  2.19s/it]\n",
      "Loss: 0.01428:  24%|██▍       | 60/250 [02:14<06:55,  2.19s/it]\n",
      "Loss: 0.01331:  24%|██▍       | 60/250 [02:14<06:55,  2.19s/it]\n",
      "Loss: 0.01411:  24%|██▍       | 60/250 [02:14<06:55,  2.19s/it]\n",
      "Loss: 0.01477:  24%|██▍       | 60/250 [02:14<06:55,  2.19s/it]\n",
      "Loss: 0.01463:  24%|██▍       | 60/250 [02:14<06:55,  2.19s/it]\n",
      "Loss: 0.01515:  24%|██▍       | 60/250 [02:14<06:55,  2.19s/it]\n",
      "Loss: 0.01231:  24%|██▍       | 60/250 [02:14<06:55,  2.19s/it]\n",
      "Loss: 0.01473:  24%|██▍       | 60/250 [02:14<06:55,  2.19s/it]\n",
      "Loss: 0.01362:  24%|██▍       | 60/250 [02:14<06:55,  2.19s/it]\n",
      "Loss: 0.01321:  24%|██▍       | 60/250 [02:15<06:55,  2.19s/it]\n",
      "Loss: 0.01391:  24%|██▍       | 60/250 [02:15<06:55,  2.19s/it]\n",
      "Loss: 0.01475:  24%|██▍       | 60/250 [02:15<06:55,  2.19s/it]\n",
      "Loss: 0.01475:  24%|██▍       | 61/250 [02:15<06:52,  2.18s/it]\n",
      "Loss: 0.01383:  24%|██▍       | 61/250 [02:15<06:52,  2.18s/it]\n",
      "Loss: 0.01434:  24%|██▍       | 61/250 [02:15<06:52,  2.18s/it]\n",
      "Loss: 0.01482:  24%|██▍       | 61/250 [02:15<06:52,  2.18s/it]\n",
      "Loss: 0.01627:  24%|██▍       | 61/250 [02:15<06:52,  2.18s/it]\n",
      "Loss: 0.01547:  24%|██▍       | 61/250 [02:15<06:52,  2.18s/it]\n",
      "Loss: 0.01407:  24%|██▍       | 61/250 [02:15<06:52,  2.18s/it]\n",
      "Loss: 0.01593:  24%|██▍       | 61/250 [02:15<06:52,  2.18s/it]\n",
      "Loss: 0.01356:  24%|██▍       | 61/250 [02:16<06:52,  2.18s/it]\n",
      "Loss: 0.01633:  24%|██▍       | 61/250 [02:16<06:52,  2.18s/it]\n",
      "Loss: 0.01561:  24%|██▍       | 61/250 [02:16<06:52,  2.18s/it]\n",
      "Loss: 0.01365:  24%|██▍       | 61/250 [02:16<06:52,  2.18s/it]\n",
      "Loss: 0.01352:  24%|██▍       | 61/250 [02:16<06:52,  2.18s/it]\n",
      "Loss: 0.01417:  24%|██▍       | 61/250 [02:16<06:52,  2.18s/it]\n",
      "Loss: 0.01260:  24%|██▍       | 61/250 [02:16<06:52,  2.18s/it]\n",
      "Loss: 0.01296:  24%|██▍       | 61/250 [02:16<06:52,  2.18s/it]\n",
      "Loss: 0.01606:  24%|██▍       | 61/250 [02:16<06:52,  2.18s/it]\n",
      "Loss: 0.01300:  24%|██▍       | 61/250 [02:17<06:52,  2.18s/it]\n",
      "Loss: 0.01408:  24%|██▍       | 61/250 [02:17<06:52,  2.18s/it]\n",
      "Loss: 0.01294:  24%|██▍       | 61/250 [02:17<06:52,  2.18s/it]\n",
      "Loss: 0.01381:  24%|██▍       | 61/250 [02:17<06:52,  2.18s/it]\n",
      "Loss: 0.01452:  24%|██▍       | 61/250 [02:17<06:52,  2.18s/it]\n",
      "Loss: 0.01452:  25%|██▍       | 62/250 [02:17<06:50,  2.18s/it]\n",
      "Loss: 0.01416:  25%|██▍       | 62/250 [02:17<06:50,  2.18s/it]\n",
      "Loss: 0.01415:  25%|██▍       | 62/250 [02:17<06:50,  2.18s/it]\n",
      "Loss: 0.01622:  25%|██▍       | 62/250 [02:17<06:50,  2.18s/it]\n",
      "Loss: 0.01322:  25%|██▍       | 62/250 [02:17<06:50,  2.18s/it]\n",
      "Loss: 0.01216:  25%|██▍       | 62/250 [02:17<06:50,  2.18s/it]\n",
      "Loss: 0.01506:  25%|██▍       | 62/250 [02:18<06:50,  2.18s/it]\n",
      "Loss: 0.01217:  25%|██▍       | 62/250 [02:18<06:50,  2.18s/it]\n",
      "Loss: 0.01883:  25%|██▍       | 62/250 [02:18<06:50,  2.18s/it]\n",
      "Loss: 0.01180:  25%|██▍       | 62/250 [02:18<06:50,  2.18s/it]\n",
      "Loss: 0.01421:  25%|██▍       | 62/250 [02:18<06:50,  2.18s/it]\n",
      "Loss: 0.01364:  25%|██▍       | 62/250 [02:18<06:50,  2.18s/it]\n",
      "Loss: 0.01387:  25%|██▍       | 62/250 [02:18<06:50,  2.18s/it]\n",
      "Loss: 0.01478:  25%|██▍       | 62/250 [02:18<06:50,  2.18s/it]\n",
      "Loss: 0.01462:  25%|██▍       | 62/250 [02:18<06:50,  2.18s/it]\n",
      "Loss: 0.01568:  25%|██▍       | 62/250 [02:19<06:50,  2.18s/it]\n",
      "Loss: 0.01387:  25%|██▍       | 62/250 [02:19<06:50,  2.18s/it]\n",
      "Loss: 0.01541:  25%|██▍       | 62/250 [02:19<06:50,  2.18s/it]\n",
      "Loss: 0.01471:  25%|██▍       | 62/250 [02:19<06:50,  2.18s/it]\n",
      "Loss: 0.01425:  25%|██▍       | 62/250 [02:19<06:50,  2.18s/it]\n",
      "Loss: 0.01355:  25%|██▍       | 62/250 [02:19<06:50,  2.18s/it]\n",
      "Loss: 0.01483:  25%|██▍       | 62/250 [02:19<06:50,  2.18s/it]\n",
      "Loss: 0.01483:  25%|██▌       | 63/250 [02:19<06:48,  2.19s/it]\n",
      "Loss: 0.01437:  25%|██▌       | 63/250 [02:19<06:48,  2.19s/it]\n",
      "Loss: 0.01399:  25%|██▌       | 63/250 [02:19<06:48,  2.19s/it]\n",
      "Loss: 0.01428:  25%|██▌       | 63/250 [02:19<06:48,  2.19s/it]\n",
      "Loss: 0.01609:  25%|██▌       | 63/250 [02:20<06:48,  2.19s/it]\n",
      "Loss: 0.01508:  25%|██▌       | 63/250 [02:20<06:48,  2.19s/it]\n",
      "Loss: 0.01415:  25%|██▌       | 63/250 [02:20<06:48,  2.19s/it]\n",
      "Loss: 0.01525:  25%|██▌       | 63/250 [02:20<06:48,  2.19s/it]\n",
      "Loss: 0.01443:  25%|██▌       | 63/250 [02:20<06:48,  2.19s/it]\n",
      "Loss: 0.01285:  25%|██▌       | 63/250 [02:20<06:48,  2.19s/it]\n",
      "Loss: 0.01608:  25%|██▌       | 63/250 [02:20<06:48,  2.19s/it]\n",
      "Loss: 0.01445:  25%|██▌       | 63/250 [02:20<06:48,  2.19s/it]\n",
      "Loss: 0.01488:  25%|██▌       | 63/250 [02:20<06:48,  2.19s/it]\n",
      "Loss: 0.01392:  25%|██▌       | 63/250 [02:21<06:48,  2.19s/it]\n",
      "Loss: 0.01286:  25%|██▌       | 63/250 [02:21<06:48,  2.19s/it]\n",
      "Loss: 0.01329:  25%|██▌       | 63/250 [02:21<06:48,  2.19s/it]\n",
      "Loss: 0.01192:  25%|██▌       | 63/250 [02:21<06:48,  2.19s/it]\n",
      "Loss: 0.01475:  25%|██▌       | 63/250 [02:21<06:48,  2.19s/it]\n",
      "Loss: 0.01524:  25%|██▌       | 63/250 [02:21<06:48,  2.19s/it]\n",
      "Loss: 0.01270:  25%|██▌       | 63/250 [02:21<06:48,  2.19s/it]\n",
      "Loss: 0.01420:  25%|██▌       | 63/250 [02:21<06:48,  2.19s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 58.09it/s]\u001b[A\n",
      "Loss: 0.01420:  26%|██▌       | 64/250 [02:21<06:46,  2.19s/it]\n",
      "Loss: 0.01362:  26%|██▌       | 64/250 [02:21<06:46,  2.19s/it]\n",
      "Loss: 0.01447:  26%|██▌       | 64/250 [02:22<06:46,  2.19s/it]\n",
      "Loss: 0.01575:  26%|██▌       | 64/250 [02:22<06:46,  2.19s/it]\n",
      "Loss: 0.01300:  26%|██▌       | 64/250 [02:22<06:46,  2.19s/it]\n",
      "Loss: 0.01212:  26%|██▌       | 64/250 [02:22<06:46,  2.19s/it]\n",
      "Loss: 0.01481:  26%|██▌       | 64/250 [02:22<06:46,  2.19s/it]\n",
      "Loss: 0.01183:  26%|██▌       | 64/250 [02:22<06:46,  2.19s/it]\n",
      "Loss: 0.01820:  26%|██▌       | 64/250 [02:22<06:46,  2.19s/it]\n",
      "Loss: 0.01184:  26%|██▌       | 64/250 [02:22<06:46,  2.19s/it]\n",
      "Loss: 0.01598:  26%|██▌       | 64/250 [02:22<06:46,  2.19s/it]\n",
      "Loss: 0.01347:  26%|██▌       | 64/250 [02:22<06:46,  2.19s/it]\n",
      "Loss: 0.01365:  26%|██▌       | 64/250 [02:23<06:46,  2.19s/it]\n",
      "Loss: 0.01478:  26%|██▌       | 64/250 [02:23<06:46,  2.19s/it]\n",
      "Loss: 0.01257:  26%|██▌       | 64/250 [02:23<06:46,  2.19s/it]\n",
      "Loss: 0.01334:  26%|██▌       | 64/250 [02:23<06:46,  2.19s/it]\n",
      "Loss: 0.01632:  26%|██▌       | 64/250 [02:23<06:46,  2.19s/it]\n",
      "Loss: 0.01300:  26%|██▌       | 64/250 [02:23<06:46,  2.19s/it]\n",
      "Loss: 0.01501:  26%|██▌       | 64/250 [02:23<06:46,  2.19s/it]\n",
      "Loss: 0.01523:  26%|██▌       | 64/250 [02:23<06:46,  2.19s/it]\n",
      "Loss: 0.01477:  26%|██▌       | 64/250 [02:23<06:46,  2.19s/it]\n",
      "Loss: 0.01494:  26%|██▌       | 64/250 [02:24<06:46,  2.19s/it]\n",
      "Loss: 0.01494:  26%|██▌       | 65/250 [02:24<06:46,  2.20s/it]\n",
      "Loss: 0.01339:  26%|██▌       | 65/250 [02:24<06:46,  2.20s/it]\n",
      "Loss: 0.01382:  26%|██▌       | 65/250 [02:24<06:46,  2.20s/it]\n",
      "Loss: 0.01342:  26%|██▌       | 65/250 [02:24<06:46,  2.20s/it]\n",
      "Loss: 0.01372:  26%|██▌       | 65/250 [02:24<06:46,  2.20s/it]\n",
      "Loss: 0.01180:  26%|██▌       | 65/250 [02:24<06:46,  2.20s/it]\n",
      "Loss: 0.01255:  26%|██▌       | 65/250 [02:24<06:46,  2.20s/it]\n",
      "Loss: 0.01317:  26%|██▌       | 65/250 [02:24<06:46,  2.20s/it]\n",
      "Loss: 0.01440:  26%|██▌       | 65/250 [02:24<06:46,  2.20s/it]\n",
      "Loss: 0.01256:  26%|██▌       | 65/250 [02:25<06:46,  2.20s/it]\n",
      "Loss: 0.01597:  26%|██▌       | 65/250 [02:25<06:46,  2.20s/it]\n",
      "Loss: 0.01427:  26%|██▌       | 65/250 [02:25<06:46,  2.20s/it]\n",
      "Loss: 0.01483:  26%|██▌       | 65/250 [02:25<06:46,  2.20s/it]\n",
      "Loss: 0.01469:  26%|██▌       | 65/250 [02:25<06:46,  2.20s/it]\n",
      "Loss: 0.01463:  26%|██▌       | 65/250 [02:25<06:46,  2.20s/it]\n",
      "Loss: 0.01530:  26%|██▌       | 65/250 [02:25<06:46,  2.20s/it]\n",
      "Loss: 0.01402:  26%|██▌       | 65/250 [02:25<06:46,  2.20s/it]\n",
      "Loss: 0.01514:  26%|██▌       | 65/250 [02:25<06:46,  2.20s/it]\n",
      "Loss: 0.01364:  26%|██▌       | 65/250 [02:25<06:46,  2.20s/it]\n",
      "Loss: 0.01295:  26%|██▌       | 65/250 [02:26<06:46,  2.20s/it]\n",
      "Loss: 0.01384:  26%|██▌       | 65/250 [02:26<06:46,  2.20s/it]\n",
      "Loss: 0.01457:  26%|██▌       | 65/250 [02:26<06:46,  2.20s/it]\n",
      "Loss: 0.01457:  26%|██▋       | 66/250 [02:26<06:44,  2.20s/it]\n",
      "Loss: 0.01363:  26%|██▋       | 66/250 [02:26<06:44,  2.20s/it]\n",
      "Loss: 0.01394:  26%|██▋       | 66/250 [02:26<06:44,  2.20s/it]\n",
      "Loss: 0.01615:  26%|██▋       | 66/250 [02:26<06:44,  2.20s/it]\n",
      "Loss: 0.01264:  26%|██▋       | 66/250 [02:26<06:44,  2.20s/it]\n",
      "Loss: 0.01194:  26%|██▋       | 66/250 [02:26<06:44,  2.20s/it]\n",
      "Loss: 0.01481:  26%|██▋       | 66/250 [02:26<06:44,  2.20s/it]\n",
      "Loss: 0.01496:  26%|██▋       | 66/250 [02:26<06:44,  2.20s/it]\n",
      "Loss: 0.01387:  26%|██▋       | 66/250 [02:27<06:44,  2.20s/it]\n",
      "Loss: 0.01623:  26%|██▋       | 66/250 [02:27<06:44,  2.20s/it]\n",
      "Loss: 0.01423:  26%|██▋       | 66/250 [02:27<06:44,  2.20s/it]\n",
      "Loss: 0.01318:  26%|██▋       | 66/250 [02:27<06:44,  2.20s/it]\n",
      "Loss: 0.01351:  26%|██▋       | 66/250 [02:27<06:44,  2.20s/it]\n",
      "Loss: 0.01481:  26%|██▋       | 66/250 [02:27<06:44,  2.20s/it]\n",
      "Loss: 0.01230:  26%|██▋       | 66/250 [02:27<06:44,  2.20s/it]\n",
      "Loss: 0.01454:  26%|██▋       | 66/250 [02:27<06:44,  2.20s/it]\n",
      "Loss: 0.01221:  26%|██▋       | 66/250 [02:27<06:44,  2.20s/it]\n",
      "Loss: 0.01458:  26%|██▋       | 66/250 [02:28<06:44,  2.20s/it]\n",
      "Loss: 0.01741:  26%|██▋       | 66/250 [02:28<06:44,  2.20s/it]\n",
      "Loss: 0.01335:  26%|██▋       | 66/250 [02:28<06:44,  2.20s/it]\n",
      "Loss: 0.01517:  26%|██▋       | 66/250 [02:28<06:44,  2.20s/it]\n",
      "Loss: 0.01517:  27%|██▋       | 67/250 [02:28<06:40,  2.19s/it]\n",
      "Loss: 0.01339:  27%|██▋       | 67/250 [02:28<06:40,  2.19s/it]\n",
      "Loss: 0.01384:  27%|██▋       | 67/250 [02:28<06:40,  2.19s/it]\n",
      "Loss: 0.01592:  27%|██▋       | 67/250 [02:28<06:40,  2.19s/it]\n",
      "Loss: 0.01237:  27%|██▋       | 67/250 [02:28<06:40,  2.19s/it]\n",
      "Loss: 0.01212:  27%|██▋       | 67/250 [02:28<06:40,  2.19s/it]\n",
      "Loss: 0.01465:  27%|██▋       | 67/250 [02:29<06:40,  2.19s/it]\n",
      "Loss: 0.01167:  27%|██▋       | 67/250 [02:29<06:40,  2.19s/it]\n",
      "Loss: 0.01805:  27%|██▋       | 67/250 [02:29<06:40,  2.19s/it]\n",
      "Loss: 0.01168:  27%|██▋       | 67/250 [02:29<06:40,  2.19s/it]\n",
      "Loss: 0.01590:  27%|██▋       | 67/250 [02:29<06:40,  2.19s/it]\n",
      "Loss: 0.01602:  27%|██▋       | 67/250 [02:29<06:40,  2.19s/it]\n",
      "Loss: 0.01396:  27%|██▋       | 67/250 [02:29<06:40,  2.19s/it]\n",
      "Loss: 0.01495:  27%|██▋       | 67/250 [02:29<06:40,  2.19s/it]\n",
      "Loss: 0.01547:  27%|██▋       | 67/250 [02:29<06:40,  2.19s/it]\n",
      "Loss: 0.01384:  27%|██▋       | 67/250 [02:29<06:40,  2.19s/it]\n",
      "Loss: 0.01235:  27%|██▋       | 67/250 [02:30<06:40,  2.19s/it]\n",
      "Loss: 0.01356:  27%|██▋       | 67/250 [02:30<06:40,  2.19s/it]\n",
      "Loss: 0.01316:  27%|██▋       | 67/250 [02:30<06:40,  2.19s/it]\n",
      "Loss: 0.01725:  27%|██▋       | 67/250 [02:30<06:40,  2.19s/it]\n",
      "Loss: 0.01320:  27%|██▋       | 67/250 [02:30<06:40,  2.19s/it]\n",
      "Loss: 0.01430:  27%|██▋       | 67/250 [02:30<06:40,  2.19s/it]\n",
      "Loss: 0.01430:  27%|██▋       | 68/250 [02:30<06:39,  2.19s/it]\n",
      "Loss: 0.01321:  27%|██▋       | 68/250 [02:30<06:39,  2.19s/it]\n",
      "Loss: 0.01431:  27%|██▋       | 68/250 [02:30<06:39,  2.19s/it]\n",
      "Loss: 0.01588:  27%|██▋       | 68/250 [02:30<06:39,  2.19s/it]\n",
      "Loss: 0.01237:  27%|██▋       | 68/250 [02:31<06:39,  2.19s/it]\n",
      "Loss: 0.01178:  27%|██▋       | 68/250 [02:31<06:39,  2.19s/it]\n",
      "Loss: 0.01475:  27%|██▋       | 68/250 [02:31<06:39,  2.19s/it]\n",
      "Loss: 0.01172:  27%|██▋       | 68/250 [02:31<06:39,  2.19s/it]\n",
      "Loss: 0.01766:  27%|██▋       | 68/250 [02:31<06:39,  2.19s/it]\n",
      "Loss: 0.01161:  27%|██▋       | 68/250 [02:31<06:39,  2.19s/it]\n",
      "Loss: 0.01588:  27%|██▋       | 68/250 [02:31<06:39,  2.19s/it]\n",
      "Loss: 0.01599:  27%|██▋       | 68/250 [02:31<06:39,  2.19s/it]\n",
      "Loss: 0.01388:  27%|██▋       | 68/250 [02:31<06:39,  2.19s/it]\n",
      "Loss: 0.01499:  27%|██▋       | 68/250 [02:31<06:39,  2.19s/it]\n",
      "Loss: 0.01522:  27%|██▋       | 68/250 [02:32<06:39,  2.19s/it]\n",
      "Loss: 0.01396:  27%|██▋       | 68/250 [02:32<06:39,  2.19s/it]\n",
      "Loss: 0.01252:  27%|██▋       | 68/250 [02:32<06:39,  2.19s/it]\n",
      "Loss: 0.01377:  27%|██▋       | 68/250 [02:32<06:39,  2.19s/it]\n",
      "Loss: 0.01374:  27%|██▋       | 68/250 [02:32<06:39,  2.19s/it]\n",
      "Loss: 0.01803:  27%|██▋       | 68/250 [02:32<06:39,  2.19s/it]\n",
      "Loss: 0.01323:  27%|██▋       | 68/250 [02:32<06:39,  2.19s/it]\n",
      "Loss: 0.01432:  27%|██▋       | 68/250 [02:32<06:39,  2.19s/it]\n",
      "Loss: 0.01432:  28%|██▊       | 69/250 [02:32<06:39,  2.21s/it]\n",
      "Loss: 0.01379:  28%|██▊       | 69/250 [02:32<06:39,  2.21s/it]\n",
      "Loss: 0.01438:  28%|██▊       | 69/250 [02:33<06:39,  2.21s/it]\n",
      "Loss: 0.01564:  28%|██▊       | 69/250 [02:33<06:39,  2.21s/it]\n",
      "Loss: 0.01245:  28%|██▊       | 69/250 [02:33<06:39,  2.21s/it]\n",
      "Loss: 0.01210:  28%|██▊       | 69/250 [02:33<06:39,  2.21s/it]\n",
      "Loss: 0.01442:  28%|██▊       | 69/250 [02:33<06:39,  2.21s/it]\n",
      "Loss: 0.01163:  28%|██▊       | 69/250 [02:33<06:39,  2.21s/it]\n",
      "Loss: 0.01778:  28%|██▊       | 69/250 [02:33<06:39,  2.21s/it]\n",
      "Loss: 0.01171:  28%|██▊       | 69/250 [02:33<06:39,  2.21s/it]\n",
      "Loss: 0.01552:  28%|██▊       | 69/250 [02:33<06:39,  2.21s/it]\n",
      "Loss: 0.01556:  28%|██▊       | 69/250 [02:33<06:39,  2.21s/it]\n",
      "Loss: 0.01371:  28%|██▊       | 69/250 [02:34<06:39,  2.21s/it]\n",
      "Loss: 0.01505:  28%|██▊       | 69/250 [02:34<06:39,  2.21s/it]\n",
      "Loss: 0.01513:  28%|██▊       | 69/250 [02:34<06:39,  2.21s/it]\n",
      "Loss: 0.01380:  28%|██▊       | 69/250 [02:34<06:39,  2.21s/it]\n",
      "Loss: 0.01240:  28%|██▊       | 69/250 [02:34<06:39,  2.21s/it]\n",
      "Loss: 0.01348:  28%|██▊       | 69/250 [02:34<06:39,  2.21s/it]\n",
      "Loss: 0.01374:  28%|██▊       | 69/250 [02:34<06:39,  2.21s/it]\n",
      "Loss: 0.01738:  28%|██▊       | 69/250 [02:34<06:39,  2.21s/it]\n",
      "Loss: 0.01327:  28%|██▊       | 69/250 [02:34<06:39,  2.21s/it]\n",
      "Loss: 0.01410:  28%|██▊       | 69/250 [02:35<06:39,  2.21s/it]\n",
      "Loss: 0.01410:  28%|██▊       | 70/250 [02:35<06:37,  2.21s/it]\n",
      "Loss: 0.01341:  28%|██▊       | 70/250 [02:35<06:37,  2.21s/it]\n",
      "Loss: 0.01364:  28%|██▊       | 70/250 [02:35<06:37,  2.21s/it]\n",
      "Loss: 0.01571:  28%|██▊       | 70/250 [02:35<06:37,  2.21s/it]\n",
      "Loss: 0.01545:  28%|██▊       | 70/250 [02:35<06:37,  2.21s/it]\n",
      "Loss: 0.01180:  28%|██▊       | 70/250 [02:35<06:37,  2.21s/it]\n",
      "Loss: 0.01287:  28%|██▊       | 70/250 [02:35<06:37,  2.21s/it]\n",
      "Loss: 0.01197:  28%|██▊       | 70/250 [02:35<06:37,  2.21s/it]\n",
      "Loss: 0.01408:  28%|██▊       | 70/250 [02:35<06:37,  2.21s/it]\n",
      "Loss: 0.01256:  28%|██▊       | 70/250 [02:36<06:37,  2.21s/it]\n",
      "Loss: 0.01585:  28%|██▊       | 70/250 [02:36<06:37,  2.21s/it]\n",
      "Loss: 0.01383:  28%|██▊       | 70/250 [02:36<06:37,  2.21s/it]\n",
      "Loss: 0.01503:  28%|██▊       | 70/250 [02:36<06:37,  2.21s/it]\n",
      "Loss: 0.01434:  28%|██▊       | 70/250 [02:36<06:37,  2.21s/it]\n",
      "Loss: 0.01451:  28%|██▊       | 70/250 [02:36<06:37,  2.21s/it]\n",
      "Loss: 0.01449:  28%|██▊       | 70/250 [02:36<06:37,  2.21s/it]\n",
      "Loss: 0.01187:  28%|██▊       | 70/250 [02:36<06:37,  2.21s/it]\n",
      "Loss: 0.01423:  28%|██▊       | 70/250 [02:36<06:37,  2.21s/it]\n",
      "Loss: 0.01360:  28%|██▊       | 70/250 [02:36<06:37,  2.21s/it]\n",
      "Loss: 0.01268:  28%|██▊       | 70/250 [02:37<06:37,  2.21s/it]\n",
      "Loss: 0.01305:  28%|██▊       | 70/250 [02:37<06:37,  2.21s/it]\n",
      "Loss: 0.01387:  28%|██▊       | 70/250 [02:37<06:37,  2.21s/it]\n",
      "Loss: 0.01387:  28%|██▊       | 71/250 [02:37<06:33,  2.20s/it]\n",
      "Loss: 0.01329:  28%|██▊       | 71/250 [02:37<06:33,  2.20s/it]\n",
      "Loss: 0.01382:  28%|██▊       | 71/250 [02:37<06:33,  2.20s/it]\n",
      "Loss: 0.01586:  28%|██▊       | 71/250 [02:37<06:33,  2.20s/it]\n",
      "Loss: 0.01253:  28%|██▊       | 71/250 [02:37<06:33,  2.20s/it]\n",
      "Loss: 0.01172:  28%|██▊       | 71/250 [02:37<06:33,  2.20s/it]\n",
      "Loss: 0.01445:  28%|██▊       | 71/250 [02:37<06:33,  2.20s/it]\n",
      "Loss: 0.01149:  28%|██▊       | 71/250 [02:37<06:33,  2.20s/it]\n",
      "Loss: 0.01849:  28%|██▊       | 71/250 [02:38<06:33,  2.20s/it]\n",
      "Loss: 0.01152:  28%|██▊       | 71/250 [02:38<06:33,  2.20s/it]\n",
      "Loss: 0.01565:  28%|██▊       | 71/250 [02:38<06:33,  2.20s/it]\n",
      "Loss: 0.01558:  28%|██▊       | 71/250 [02:38<06:33,  2.20s/it]\n",
      "Loss: 0.01366:  28%|██▊       | 71/250 [02:38<06:33,  2.20s/it]\n",
      "Loss: 0.01477:  28%|██▊       | 71/250 [02:38<06:33,  2.20s/it]\n",
      "Loss: 0.01527:  28%|██▊       | 71/250 [02:38<06:33,  2.20s/it]\n",
      "Loss: 0.01390:  28%|██▊       | 71/250 [02:38<06:33,  2.20s/it]\n",
      "Loss: 0.01261:  28%|██▊       | 71/250 [02:38<06:33,  2.20s/it]\n",
      "Loss: 0.01343:  28%|██▊       | 71/250 [02:39<06:33,  2.20s/it]\n",
      "Loss: 0.01350:  28%|██▊       | 71/250 [02:39<06:33,  2.20s/it]\n",
      "Loss: 0.01774:  28%|██▊       | 71/250 [02:39<06:33,  2.20s/it]\n",
      "Loss: 0.01291:  28%|██▊       | 71/250 [02:39<06:33,  2.20s/it]\n",
      "Loss: 0.01405:  28%|██▊       | 71/250 [02:39<06:33,  2.20s/it]\n",
      "Loss: 0.01405:  29%|██▉       | 72/250 [02:39<06:31,  2.20s/it]\n",
      "Loss: 0.01417:  29%|██▉       | 72/250 [02:39<06:31,  2.20s/it]\n",
      "Loss: 0.01370:  29%|██▉       | 72/250 [02:39<06:31,  2.20s/it]\n",
      "Loss: 0.01406:  29%|██▉       | 72/250 [02:39<06:31,  2.20s/it]\n",
      "Loss: 0.01568:  29%|██▉       | 72/250 [02:39<06:31,  2.20s/it]\n",
      "Loss: 0.01530:  29%|██▉       | 72/250 [02:39<06:31,  2.20s/it]\n",
      "Loss: 0.01380:  29%|██▉       | 72/250 [02:40<06:31,  2.20s/it]\n",
      "Loss: 0.01495:  29%|██▉       | 72/250 [02:40<06:31,  2.20s/it]\n",
      "Loss: 0.01297:  29%|██▉       | 72/250 [02:40<06:31,  2.20s/it]\n",
      "Loss: 0.01585:  29%|██▉       | 72/250 [02:40<06:31,  2.20s/it]\n",
      "Loss: 0.01406:  29%|██▉       | 72/250 [02:40<06:31,  2.20s/it]\n",
      "Loss: 0.01332:  29%|██▉       | 72/250 [02:40<06:31,  2.20s/it]\n",
      "Loss: 0.01329:  29%|██▉       | 72/250 [02:40<06:31,  2.20s/it]\n",
      "Loss: 0.01364:  29%|██▉       | 72/250 [02:40<06:31,  2.20s/it]\n",
      "Loss: 0.01217:  29%|██▉       | 72/250 [02:40<06:31,  2.20s/it]\n",
      "Loss: 0.01304:  29%|██▉       | 72/250 [02:41<06:31,  2.20s/it]\n",
      "Loss: 0.01536:  29%|██▉       | 72/250 [02:41<06:31,  2.20s/it]\n",
      "Loss: 0.01278:  29%|██▉       | 72/250 [02:41<06:31,  2.20s/it]\n",
      "Loss: 0.01478:  29%|██▉       | 72/250 [02:41<06:31,  2.20s/it]\n",
      "Loss: 0.01473:  29%|██▉       | 72/250 [02:41<06:31,  2.20s/it]\n",
      "Loss: 0.01371:  29%|██▉       | 72/250 [02:41<06:31,  2.20s/it]\n",
      "Loss: 0.01388:  29%|██▉       | 72/250 [02:41<06:31,  2.20s/it]\n",
      "Loss: 0.01388:  29%|██▉       | 73/250 [02:41<06:28,  2.20s/it]\n",
      "Loss: 0.01315:  29%|██▉       | 73/250 [02:41<06:28,  2.20s/it]\n",
      "Loss: 0.01493:  29%|██▉       | 73/250 [02:41<06:28,  2.20s/it]\n",
      "Loss: 0.01578:  29%|██▉       | 73/250 [02:41<06:28,  2.20s/it]\n",
      "Loss: 0.01241:  29%|██▉       | 73/250 [02:42<06:28,  2.20s/it]\n",
      "Loss: 0.01176:  29%|██▉       | 73/250 [02:42<06:28,  2.20s/it]\n",
      "Loss: 0.01386:  29%|██▉       | 73/250 [02:42<06:28,  2.20s/it]\n",
      "Loss: 0.01452:  29%|██▉       | 73/250 [02:42<06:28,  2.20s/it]\n",
      "Loss: 0.01328:  29%|██▉       | 73/250 [02:42<06:28,  2.20s/it]\n",
      "Loss: 0.01593:  29%|██▉       | 73/250 [02:42<06:28,  2.20s/it]\n",
      "Loss: 0.01398:  29%|██▉       | 73/250 [02:42<06:28,  2.20s/it]\n",
      "Loss: 0.01333:  29%|██▉       | 73/250 [02:42<06:28,  2.20s/it]\n",
      "Loss: 0.01331:  29%|██▉       | 73/250 [02:42<06:28,  2.20s/it]\n",
      "Loss: 0.01397:  29%|██▉       | 73/250 [02:42<06:28,  2.20s/it]\n",
      "Loss: 0.01235:  29%|██▉       | 73/250 [02:43<06:28,  2.20s/it]\n",
      "Loss: 0.01250:  29%|██▉       | 73/250 [02:43<06:28,  2.20s/it]\n",
      "Loss: 0.01516:  29%|██▉       | 73/250 [02:43<06:28,  2.20s/it]\n",
      "Loss: 0.01247:  29%|██▉       | 73/250 [02:43<06:28,  2.20s/it]\n",
      "Loss: 0.01543:  29%|██▉       | 73/250 [02:43<06:28,  2.20s/it]\n",
      "Loss: 0.01481:  29%|██▉       | 73/250 [02:43<06:28,  2.20s/it]\n",
      "Loss: 0.01376:  29%|██▉       | 73/250 [02:43<06:28,  2.20s/it]\n",
      "Loss: 0.01418:  29%|██▉       | 73/250 [02:43<06:28,  2.20s/it]\n",
      "Loss: 0.01418:  30%|██▉       | 74/250 [02:43<06:26,  2.19s/it]\n",
      "Loss: 0.01355:  30%|██▉       | 74/250 [02:43<06:26,  2.19s/it]\n",
      "Loss: 0.01361:  30%|██▉       | 74/250 [02:44<06:26,  2.19s/it]\n",
      "Loss: 0.01540:  30%|██▉       | 74/250 [02:44<06:26,  2.19s/it]\n",
      "Loss: 0.01231:  30%|██▉       | 74/250 [02:44<06:26,  2.19s/it]\n",
      "Loss: 0.01167:  30%|██▉       | 74/250 [02:44<06:26,  2.19s/it]\n",
      "Loss: 0.01418:  30%|██▉       | 74/250 [02:44<06:26,  2.19s/it]\n",
      "Loss: 0.01139:  30%|██▉       | 74/250 [02:44<06:26,  2.19s/it]\n",
      "Loss: 0.01754:  30%|██▉       | 74/250 [02:44<06:26,  2.19s/it]\n",
      "Loss: 0.01200:  30%|██▉       | 74/250 [02:44<06:26,  2.19s/it]\n",
      "Loss: 0.01531:  30%|██▉       | 74/250 [02:44<06:26,  2.19s/it]\n",
      "Loss: 0.01612:  30%|██▉       | 74/250 [02:44<06:26,  2.19s/it]\n",
      "Loss: 0.01359:  30%|██▉       | 74/250 [02:45<06:26,  2.19s/it]\n",
      "Loss: 0.01491:  30%|██▉       | 74/250 [02:45<06:26,  2.19s/it]\n",
      "Loss: 0.01518:  30%|██▉       | 74/250 [02:45<06:26,  2.19s/it]\n",
      "Loss: 0.01338:  30%|██▉       | 74/250 [02:45<06:26,  2.19s/it]\n",
      "Loss: 0.01262:  30%|██▉       | 74/250 [02:45<06:26,  2.19s/it]\n",
      "Loss: 0.01308:  30%|██▉       | 74/250 [02:45<06:26,  2.19s/it]\n",
      "Loss: 0.01438:  30%|██▉       | 74/250 [02:45<06:26,  2.19s/it]\n",
      "Loss: 0.01441:  30%|██▉       | 74/250 [02:45<06:26,  2.19s/it]\n",
      "Loss: 0.01354:  30%|██▉       | 74/250 [02:45<06:26,  2.19s/it]\n",
      "Loss: 0.01402:  30%|██▉       | 74/250 [02:46<06:26,  2.19s/it]\n",
      "Loss: 0.01402:  30%|███       | 75/250 [02:46<06:25,  2.20s/it]\n",
      "Loss: 0.01329:  30%|███       | 75/250 [02:46<06:25,  2.20s/it]\n",
      "Loss: 0.01351:  30%|███       | 75/250 [02:46<06:25,  2.20s/it]\n",
      "Loss: 0.01588:  30%|███       | 75/250 [02:46<06:25,  2.20s/it]\n",
      "Loss: 0.01222:  30%|███       | 75/250 [02:46<06:25,  2.20s/it]\n",
      "Loss: 0.01147:  30%|███       | 75/250 [02:46<06:25,  2.20s/it]\n",
      "Loss: 0.01431:  30%|███       | 75/250 [02:46<06:25,  2.20s/it]\n",
      "Loss: 0.01215:  30%|███       | 75/250 [02:46<06:25,  2.20s/it]\n",
      "Loss: 0.01746:  30%|███       | 75/250 [02:46<06:25,  2.20s/it]\n",
      "Loss: 0.01170:  30%|███       | 75/250 [02:46<06:25,  2.20s/it]\n",
      "Loss: 0.01368:  30%|███       | 75/250 [02:47<06:25,  2.20s/it]\n",
      "Loss: 0.01298:  30%|███       | 75/250 [02:47<06:25,  2.20s/it]\n",
      "Loss: 0.01307:  30%|███       | 75/250 [02:47<06:25,  2.20s/it]\n",
      "Loss: 0.01346:  30%|███       | 75/250 [02:47<06:25,  2.20s/it]\n",
      "Loss: 0.01368:  30%|███       | 75/250 [02:47<06:25,  2.20s/it]\n",
      "Loss: 0.01505:  30%|███       | 75/250 [02:47<06:25,  2.20s/it]\n",
      "Loss: 0.01363:  30%|███       | 75/250 [02:47<06:25,  2.20s/it]\n",
      "Loss: 0.01585:  30%|███       | 75/250 [02:47<06:25,  2.20s/it]\n",
      "Loss: 0.01334:  30%|███       | 75/250 [02:47<06:25,  2.20s/it]\n",
      "Loss: 0.01406:  30%|███       | 75/250 [02:48<06:25,  2.20s/it]\n",
      "Loss: 0.01285:  30%|███       | 75/250 [02:48<06:25,  2.20s/it]\n",
      "Loss: 0.01409:  30%|███       | 75/250 [02:48<06:25,  2.20s/it]\n",
      "Loss: 0.01409:  30%|███       | 76/250 [02:48<06:22,  2.20s/it]\n",
      "Loss: 0.01357:  30%|███       | 76/250 [02:48<06:22,  2.20s/it]\n",
      "Loss: 0.01369:  30%|███       | 76/250 [02:48<06:22,  2.20s/it]\n",
      "Loss: 0.01532:  30%|███       | 76/250 [02:48<06:22,  2.20s/it]\n",
      "Loss: 0.01229:  30%|███       | 76/250 [02:48<06:22,  2.20s/it]\n",
      "Loss: 0.01458:  30%|███       | 76/250 [02:48<06:22,  2.20s/it]\n",
      "Loss: 0.01408:  30%|███       | 76/250 [02:48<06:22,  2.20s/it]\n",
      "Loss: 0.01451:  30%|███       | 76/250 [02:48<06:22,  2.20s/it]\n",
      "Loss: 0.01292:  30%|███       | 76/250 [02:49<06:22,  2.20s/it]\n",
      "Loss: 0.01577:  30%|███       | 76/250 [02:49<06:22,  2.20s/it]\n",
      "Loss: 0.01352:  30%|███       | 76/250 [02:49<06:22,  2.20s/it]\n",
      "Loss: 0.01322:  30%|███       | 76/250 [02:49<06:22,  2.20s/it]\n",
      "Loss: 0.01312:  30%|███       | 76/250 [02:49<06:22,  2.20s/it]\n",
      "Loss: 0.01348:  30%|███       | 76/250 [02:49<06:22,  2.20s/it]\n",
      "Loss: 0.01175:  30%|███       | 76/250 [02:49<06:22,  2.20s/it]\n",
      "Loss: 0.01300:  30%|███       | 76/250 [02:49<06:22,  2.20s/it]\n",
      "Loss: 0.01533:  30%|███       | 76/250 [02:50<06:22,  2.20s/it]\n",
      "Loss: 0.01237:  30%|███       | 76/250 [02:50<06:22,  2.20s/it]\n",
      "Loss: 0.01434:  30%|███       | 76/250 [02:50<06:22,  2.20s/it]\n",
      "Loss: 0.01462:  30%|███       | 76/250 [02:50<06:22,  2.20s/it]\n",
      "Loss: 0.01377:  30%|███       | 76/250 [02:50<06:22,  2.20s/it]\n",
      "Loss: 0.01424:  30%|███       | 76/250 [02:50<06:22,  2.20s/it]\n",
      "Loss: 0.01424:  31%|███       | 77/250 [02:50<06:25,  2.23s/it]\n",
      "Loss: 0.01333:  31%|███       | 77/250 [02:50<06:25,  2.23s/it]\n",
      "Loss: 0.01381:  31%|███       | 77/250 [02:50<06:25,  2.23s/it]\n",
      "Loss: 0.01547:  31%|███       | 77/250 [02:50<06:25,  2.23s/it]\n",
      "Loss: 0.01233:  31%|███       | 77/250 [02:50<06:25,  2.23s/it]\n",
      "Loss: 0.01143:  31%|███       | 77/250 [02:51<06:25,  2.23s/it]\n",
      "Loss: 0.01445:  31%|███       | 77/250 [02:51<06:25,  2.23s/it]\n",
      "Loss: 0.01121:  31%|███       | 77/250 [02:51<06:25,  2.23s/it]\n",
      "Loss: 0.01758:  31%|███       | 77/250 [02:51<06:25,  2.23s/it]\n",
      "Loss: 0.01137:  31%|███       | 77/250 [02:51<06:25,  2.23s/it]\n",
      "Loss: 0.01510:  31%|███       | 77/250 [02:51<06:25,  2.23s/it]\n",
      "Loss: 0.01555:  31%|███       | 77/250 [02:51<06:25,  2.23s/it]\n",
      "Loss: 0.01319:  31%|███       | 77/250 [02:51<06:25,  2.23s/it]\n",
      "Loss: 0.01443:  31%|███       | 77/250 [02:51<06:25,  2.23s/it]\n",
      "Loss: 0.01514:  31%|███       | 77/250 [02:52<06:25,  2.23s/it]\n",
      "Loss: 0.01328:  31%|███       | 77/250 [02:52<06:25,  2.23s/it]\n",
      "Loss: 0.01220:  31%|███       | 77/250 [02:52<06:25,  2.23s/it]\n",
      "Loss: 0.01321:  31%|███       | 77/250 [02:52<06:25,  2.23s/it]\n",
      "Loss: 0.01301:  31%|███       | 77/250 [02:52<06:25,  2.23s/it]\n",
      "Loss: 0.01690:  31%|███       | 77/250 [02:52<06:25,  2.23s/it]\n",
      "Loss: 0.01348:  31%|███       | 77/250 [02:52<06:25,  2.23s/it]\n",
      "Loss: 0.01388:  31%|███       | 77/250 [02:52<06:25,  2.23s/it]\n",
      "Loss: 0.01388:  31%|███       | 78/250 [02:52<06:23,  2.23s/it]\n",
      "Loss: 0.01328:  31%|███       | 78/250 [02:52<06:23,  2.23s/it]\n",
      "Loss: 0.01392:  31%|███       | 78/250 [02:52<06:23,  2.23s/it]\n",
      "Loss: 0.01506:  31%|███       | 78/250 [02:53<06:23,  2.23s/it]\n",
      "Loss: 0.01214:  31%|███       | 78/250 [02:53<06:23,  2.23s/it]\n",
      "Loss: 0.01224:  31%|███       | 78/250 [02:53<06:23,  2.23s/it]\n",
      "Loss: 0.01399:  31%|███       | 78/250 [02:53<06:23,  2.23s/it]\n",
      "Loss: 0.01121:  31%|███       | 78/250 [02:53<06:23,  2.23s/it]\n",
      "Loss: 0.01757:  31%|███       | 78/250 [02:53<06:23,  2.23s/it]\n",
      "Loss: 0.01156:  31%|███       | 78/250 [02:53<06:23,  2.23s/it]\n",
      "Loss: 0.01585:  31%|███       | 78/250 [02:53<06:23,  2.23s/it]\n",
      "Loss: 0.01554:  31%|███       | 78/250 [02:53<06:23,  2.23s/it]\n",
      "Loss: 0.01351:  31%|███       | 78/250 [02:54<06:23,  2.23s/it]\n",
      "Loss: 0.01457:  31%|███       | 78/250 [02:54<06:23,  2.23s/it]\n",
      "Loss: 0.01497:  31%|███       | 78/250 [02:54<06:23,  2.23s/it]\n",
      "Loss: 0.01322:  31%|███       | 78/250 [02:54<06:23,  2.23s/it]\n",
      "Loss: 0.01186:  31%|███       | 78/250 [02:54<06:23,  2.23s/it]\n",
      "Loss: 0.01335:  31%|███       | 78/250 [02:54<06:23,  2.23s/it]\n",
      "Loss: 0.01305:  31%|███       | 78/250 [02:54<06:23,  2.23s/it]\n",
      "Loss: 0.01671:  31%|███       | 78/250 [02:54<06:23,  2.23s/it]\n",
      "Loss: 0.01260:  31%|███       | 78/250 [02:54<06:23,  2.23s/it]\n",
      "Loss: 0.01368:  31%|███       | 78/250 [02:54<06:23,  2.23s/it]\n",
      "Loss: 0.01368:  32%|███▏      | 79/250 [02:54<06:19,  2.22s/it]\n",
      "Loss: 0.01391:  32%|███▏      | 79/250 [02:55<06:19,  2.22s/it]\n",
      "Loss: 0.01345:  32%|███▏      | 79/250 [02:55<06:19,  2.22s/it]\n",
      "Loss: 0.01500:  32%|███▏      | 79/250 [02:55<06:19,  2.22s/it]\n",
      "Loss: 0.01215:  32%|███▏      | 79/250 [02:55<06:19,  2.22s/it]\n",
      "Loss: 0.01105:  32%|███▏      | 79/250 [02:55<06:19,  2.22s/it]\n",
      "Loss: 0.01376:  32%|███▏      | 79/250 [02:55<06:19,  2.22s/it]\n",
      "Loss: 0.01446:  32%|███▏      | 79/250 [02:55<06:19,  2.22s/it]\n",
      "Loss: 0.01336:  32%|███▏      | 79/250 [02:55<06:19,  2.22s/it]\n",
      "Loss: 0.01561:  32%|███▏      | 79/250 [02:55<06:19,  2.22s/it]\n",
      "Loss: 0.01350:  32%|███▏      | 79/250 [02:56<06:19,  2.22s/it]\n",
      "Loss: 0.01340:  32%|███▏      | 79/250 [02:56<06:19,  2.22s/it]\n",
      "Loss: 0.01298:  32%|███▏      | 79/250 [02:56<06:19,  2.22s/it]\n",
      "Loss: 0.01335:  32%|███▏      | 79/250 [02:56<06:19,  2.22s/it]\n",
      "Loss: 0.01391:  32%|███▏      | 79/250 [02:56<06:19,  2.22s/it]\n",
      "Loss: 0.01478:  32%|███▏      | 79/250 [02:56<06:19,  2.22s/it]\n",
      "Loss: 0.01282:  32%|███▏      | 79/250 [02:56<06:19,  2.22s/it]\n",
      "Loss: 0.01453:  32%|███▏      | 79/250 [02:56<06:19,  2.22s/it]\n",
      "Loss: 0.01288:  32%|███▏      | 79/250 [02:56<06:19,  2.22s/it]\n",
      "Loss: 0.01365:  32%|███▏      | 79/250 [02:56<06:19,  2.22s/it]\n",
      "Loss: 0.01315:  32%|███▏      | 79/250 [02:57<06:19,  2.22s/it]\n",
      "Loss: 0.01348:  32%|███▏      | 79/250 [02:57<06:19,  2.22s/it]\n",
      "Loss: 0.01348:  32%|███▏      | 80/250 [02:57<06:17,  2.22s/it]\n",
      "Loss: 0.01361:  32%|███▏      | 80/250 [02:57<06:17,  2.22s/it]\n",
      "Loss: 0.01322:  32%|███▏      | 80/250 [02:57<06:17,  2.22s/it]\n",
      "Loss: 0.01326:  32%|███▏      | 80/250 [02:57<06:17,  2.22s/it]\n",
      "Loss: 0.01484:  32%|███▏      | 80/250 [02:57<06:17,  2.22s/it]\n",
      "Loss: 0.01470:  32%|███▏      | 80/250 [02:57<06:17,  2.22s/it]\n",
      "Loss: 0.01333:  32%|███▏      | 80/250 [02:57<06:17,  2.22s/it]\n",
      "Loss: 0.01493:  32%|███▏      | 80/250 [02:57<06:17,  2.22s/it]\n",
      "Loss: 0.01296:  32%|███▏      | 80/250 [02:58<06:17,  2.22s/it]\n",
      "Loss: 0.01545:  32%|███▏      | 80/250 [02:58<06:17,  2.22s/it]\n",
      "Loss: 0.01384:  32%|███▏      | 80/250 [02:58<06:17,  2.22s/it]\n",
      "Loss: 0.01278:  32%|███▏      | 80/250 [02:58<06:17,  2.22s/it]\n",
      "Loss: 0.01298:  32%|███▏      | 80/250 [02:58<06:17,  2.22s/it]\n",
      "Loss: 0.01339:  32%|███▏      | 80/250 [02:58<06:17,  2.22s/it]\n",
      "Loss: 0.01219:  32%|███▏      | 80/250 [02:58<06:17,  2.22s/it]\n",
      "Loss: 0.01246:  32%|███▏      | 80/250 [02:58<06:17,  2.22s/it]\n",
      "Loss: 0.01537:  32%|███▏      | 80/250 [02:58<06:17,  2.22s/it]\n",
      "Loss: 0.01231:  32%|███▏      | 80/250 [02:58<06:17,  2.22s/it]\n",
      "Loss: 0.01442:  32%|███▏      | 80/250 [02:59<06:17,  2.22s/it]\n",
      "Loss: 0.01447:  32%|███▏      | 80/250 [02:59<06:17,  2.22s/it]\n",
      "Loss: 0.01341:  32%|███▏      | 80/250 [02:59<06:17,  2.22s/it]\n",
      "Loss: 0.01387:  32%|███▏      | 80/250 [02:59<06:17,  2.22s/it]\n",
      "Loss: 0.01387:  32%|███▏      | 81/250 [02:59<06:13,  2.21s/it]\n",
      "Loss: 0.01291:  32%|███▏      | 81/250 [02:59<06:13,  2.21s/it]\n",
      "Loss: 0.01336:  32%|███▏      | 81/250 [02:59<06:13,  2.21s/it]\n",
      "Loss: 0.01513:  32%|███▏      | 81/250 [02:59<06:13,  2.21s/it]\n",
      "Loss: 0.01214:  32%|███▏      | 81/250 [02:59<06:13,  2.21s/it]\n",
      "Loss: 0.01147:  32%|███▏      | 81/250 [02:59<06:13,  2.21s/it]\n",
      "Loss: 0.01385:  32%|███▏      | 81/250 [02:59<06:13,  2.21s/it]\n",
      "Loss: 0.01112:  32%|███▏      | 81/250 [03:00<06:13,  2.21s/it]\n",
      "Loss: 0.01268:  32%|███▏      | 81/250 [03:00<06:13,  2.21s/it]\n",
      "Loss: 0.01537:  32%|███▏      | 81/250 [03:00<06:13,  2.21s/it]\n",
      "Loss: 0.01451:  32%|███▏      | 81/250 [03:00<06:13,  2.21s/it]\n",
      "Loss: 0.01279:  32%|███▏      | 81/250 [03:00<06:13,  2.21s/it]\n",
      "Loss: 0.01272:  32%|███▏      | 81/250 [03:00<06:13,  2.21s/it]\n",
      "Loss: 0.01360:  32%|███▏      | 81/250 [03:00<06:13,  2.21s/it]\n",
      "Loss: 0.01178:  32%|███▏      | 81/250 [03:00<06:13,  2.21s/it]\n",
      "Loss: 0.01229:  32%|███▏      | 81/250 [03:00<06:13,  2.21s/it]\n",
      "Loss: 0.01541:  32%|███▏      | 81/250 [03:01<06:13,  2.21s/it]\n",
      "Loss: 0.01254:  32%|███▏      | 81/250 [03:01<06:13,  2.21s/it]\n",
      "Loss: 0.01437:  32%|███▏      | 81/250 [03:01<06:13,  2.21s/it]\n",
      "Loss: 0.01438:  32%|███▏      | 81/250 [03:01<06:13,  2.21s/it]\n",
      "Loss: 0.01355:  32%|███▏      | 81/250 [03:01<06:13,  2.21s/it]\n",
      "Loss: 0.01358:  32%|███▏      | 81/250 [03:01<06:13,  2.21s/it]\n",
      "Loss: 0.01358:  33%|███▎      | 82/250 [03:01<06:11,  2.21s/it]\n",
      "Loss: 0.01287:  33%|███▎      | 82/250 [03:01<06:11,  2.21s/it]\n",
      "Loss: 0.01374:  33%|███▎      | 82/250 [03:01<06:11,  2.21s/it]\n",
      "Loss: 0.01589:  33%|███▎      | 82/250 [03:01<06:11,  2.21s/it]\n",
      "Loss: 0.01208:  33%|███▎      | 82/250 [03:01<06:11,  2.21s/it]\n",
      "Loss: 0.01152:  33%|███▎      | 82/250 [03:02<06:11,  2.21s/it]\n",
      "Loss: 0.01384:  33%|███▎      | 82/250 [03:02<06:11,  2.21s/it]\n",
      "Loss: 0.01174:  33%|███▎      | 82/250 [03:02<06:11,  2.21s/it]\n",
      "Loss: 0.01692:  33%|███▎      | 82/250 [03:02<06:11,  2.21s/it]\n",
      "Loss: 0.01142:  33%|███▎      | 82/250 [03:02<06:11,  2.21s/it]\n",
      "Loss: 0.01516:  33%|███▎      | 82/250 [03:02<06:11,  2.21s/it]\n",
      "Loss: 0.01514:  33%|███▎      | 82/250 [03:02<06:11,  2.21s/it]\n",
      "Loss: 0.01325:  33%|███▎      | 82/250 [03:02<06:11,  2.21s/it]\n",
      "Loss: 0.01415:  33%|███▎      | 82/250 [03:02<06:11,  2.21s/it]\n",
      "Loss: 0.01500:  33%|███▎      | 82/250 [03:03<06:11,  2.21s/it]\n",
      "Loss: 0.01284:  33%|███▎      | 82/250 [03:03<06:11,  2.21s/it]\n",
      "Loss: 0.01494:  33%|███▎      | 82/250 [03:03<06:11,  2.21s/it]\n",
      "Loss: 0.01226:  33%|███▎      | 82/250 [03:03<06:11,  2.21s/it]\n",
      "Loss: 0.01446:  33%|███▎      | 82/250 [03:03<06:11,  2.21s/it]\n",
      "Loss: 0.01442:  33%|███▎      | 82/250 [03:03<06:11,  2.21s/it]\n",
      "Loss: 0.01292:  33%|███▎      | 82/250 [03:03<06:11,  2.21s/it]\n",
      "Loss: 0.01376:  33%|███▎      | 82/250 [03:03<06:11,  2.21s/it]\n",
      "Loss: 0.01376:  33%|███▎      | 83/250 [03:03<06:08,  2.21s/it]\n",
      "Loss: 0.01309:  33%|███▎      | 83/250 [03:03<06:08,  2.21s/it]\n",
      "Loss: 0.01341:  33%|███▎      | 83/250 [03:03<06:08,  2.21s/it]\n",
      "Loss: 0.01503:  33%|███▎      | 83/250 [03:04<06:08,  2.21s/it]\n",
      "Loss: 0.01183:  33%|███▎      | 83/250 [03:04<06:08,  2.21s/it]\n",
      "Loss: 0.01136:  33%|███▎      | 83/250 [03:04<06:08,  2.21s/it]\n",
      "Loss: 0.01367:  33%|███▎      | 83/250 [03:04<06:08,  2.21s/it]\n",
      "Loss: 0.01083:  33%|███▎      | 83/250 [03:04<06:08,  2.21s/it]\n",
      "Loss: 0.01678:  33%|███▎      | 83/250 [03:04<06:08,  2.21s/it]\n",
      "Loss: 0.01183:  33%|███▎      | 83/250 [03:04<06:08,  2.21s/it]\n",
      "Loss: 0.01583:  33%|███▎      | 83/250 [03:04<06:08,  2.21s/it]\n",
      "Loss: 0.01538:  33%|███▎      | 83/250 [03:04<06:08,  2.21s/it]\n",
      "Loss: 0.01356:  33%|███▎      | 83/250 [03:05<06:08,  2.21s/it]\n",
      "Loss: 0.01419:  33%|███▎      | 83/250 [03:05<06:08,  2.21s/it]\n",
      "Loss: 0.01455:  33%|███▎      | 83/250 [03:05<06:08,  2.21s/it]\n",
      "Loss: 0.01289:  33%|███▎      | 83/250 [03:05<06:08,  2.21s/it]\n",
      "Loss: 0.01190:  33%|███▎      | 83/250 [03:05<06:08,  2.21s/it]\n",
      "Loss: 0.01294:  33%|███▎      | 83/250 [03:05<06:08,  2.21s/it]\n",
      "Loss: 0.01280:  33%|███▎      | 83/250 [03:05<06:08,  2.21s/it]\n",
      "Loss: 0.01718:  33%|███▎      | 83/250 [03:05<06:08,  2.21s/it]\n",
      "Loss: 0.01315:  33%|███▎      | 83/250 [03:05<06:08,  2.21s/it]\n",
      "Loss: 0.01396:  33%|███▎      | 83/250 [03:05<06:08,  2.21s/it]\n",
      "Loss: 0.01396:  34%|███▎      | 84/250 [03:05<06:07,  2.21s/it]\n",
      "Loss: 0.01282:  34%|███▎      | 84/250 [03:06<06:07,  2.21s/it]\n",
      "Loss: 0.01335:  34%|███▎      | 84/250 [03:06<06:07,  2.21s/it]\n",
      "Loss: 0.01497:  34%|███▎      | 84/250 [03:06<06:07,  2.21s/it]\n",
      "Loss: 0.01204:  34%|███▎      | 84/250 [03:06<06:07,  2.21s/it]\n",
      "Loss: 0.01095:  34%|███▎      | 84/250 [03:06<06:07,  2.21s/it]\n",
      "Loss: 0.01382:  34%|███▎      | 84/250 [03:06<06:07,  2.21s/it]\n",
      "Loss: 0.01077:  34%|███▎      | 84/250 [03:06<06:07,  2.21s/it]\n",
      "Loss: 0.01692:  34%|███▎      | 84/250 [03:06<06:07,  2.21s/it]\n",
      "Loss: 0.01138:  34%|███▎      | 84/250 [03:06<06:07,  2.21s/it]\n",
      "Loss: 0.01521:  34%|███▎      | 84/250 [03:07<06:07,  2.21s/it]\n",
      "Loss: 0.01569:  34%|███▎      | 84/250 [03:07<06:07,  2.21s/it]\n",
      "Loss: 0.01282:  34%|███▎      | 84/250 [03:07<06:07,  2.21s/it]\n",
      "Loss: 0.01338:  34%|███▎      | 84/250 [03:07<06:07,  2.21s/it]\n",
      "Loss: 0.01238:  34%|███▎      | 84/250 [03:07<06:07,  2.21s/it]\n",
      "Loss: 0.01427:  34%|███▎      | 84/250 [03:07<06:07,  2.21s/it]\n",
      "Loss: 0.01284:  34%|███▎      | 84/250 [03:07<06:07,  2.21s/it]\n",
      "Loss: 0.01496:  34%|███▎      | 84/250 [03:07<06:07,  2.21s/it]\n",
      "Loss: 0.01259:  34%|███▎      | 84/250 [03:07<06:07,  2.21s/it]\n",
      "Loss: 0.01330:  34%|███▎      | 84/250 [03:08<06:07,  2.21s/it]\n",
      "Loss: 0.01311:  34%|███▎      | 84/250 [03:08<06:07,  2.21s/it]\n",
      "Loss: 0.01336:  34%|███▎      | 84/250 [03:08<06:07,  2.21s/it]\n",
      "Loss: 0.01336:  34%|███▍      | 85/250 [03:08<06:04,  2.21s/it]\n",
      "Loss: 0.01264:  34%|███▍      | 85/250 [03:08<06:04,  2.21s/it]\n",
      "Loss: 0.01351:  34%|███▍      | 85/250 [03:08<06:04,  2.21s/it]\n",
      "Loss: 0.01499:  34%|███▍      | 85/250 [03:08<06:04,  2.21s/it]\n",
      "Loss: 0.01194:  34%|███▍      | 85/250 [03:08<06:04,  2.21s/it]\n",
      "Loss: 0.01105:  34%|███▍      | 85/250 [03:08<06:04,  2.21s/it]\n",
      "Loss: 0.01420:  34%|███▍      | 85/250 [03:08<06:04,  2.21s/it]\n",
      "Loss: 0.01117:  34%|███▍      | 85/250 [03:08<06:04,  2.21s/it]\n",
      "Loss: 0.01731:  34%|███▍      | 85/250 [03:09<06:04,  2.21s/it]\n",
      "Loss: 0.01130:  34%|███▍      | 85/250 [03:09<06:04,  2.21s/it]\n",
      "Loss: 0.01483:  34%|███▍      | 85/250 [03:09<06:04,  2.21s/it]\n",
      "Loss: 0.01575:  34%|███▍      | 85/250 [03:09<06:04,  2.21s/it]\n",
      "Loss: 0.01295:  34%|███▍      | 85/250 [03:09<06:04,  2.21s/it]\n",
      "Loss: 0.01304:  34%|███▍      | 85/250 [03:09<06:04,  2.21s/it]\n",
      "Loss: 0.01201:  34%|███▍      | 85/250 [03:09<06:04,  2.21s/it]\n",
      "Loss: 0.01240:  34%|███▍      | 85/250 [03:09<06:04,  2.21s/it]\n",
      "Loss: 0.01284:  34%|███▍      | 85/250 [03:09<06:04,  2.21s/it]\n",
      "Loss: 0.01417:  34%|███▍      | 85/250 [03:10<06:04,  2.21s/it]\n",
      "Loss: 0.01296:  34%|███▍      | 85/250 [03:10<06:04,  2.21s/it]\n",
      "Loss: 0.01375:  34%|███▍      | 85/250 [03:10<06:04,  2.21s/it]\n",
      "Loss: 0.01319:  34%|███▍      | 85/250 [03:10<06:04,  2.21s/it]\n",
      "Loss: 0.01373:  34%|███▍      | 85/250 [03:10<06:04,  2.21s/it]\n",
      "Loss: 0.01373:  34%|███▍      | 86/250 [03:10<06:01,  2.21s/it]\n",
      "Loss: 0.01288:  34%|███▍      | 86/250 [03:10<06:01,  2.21s/it]\n",
      "Loss: 0.01374:  34%|███▍      | 86/250 [03:10<06:01,  2.21s/it]\n",
      "Loss: 0.01487:  34%|███▍      | 86/250 [03:10<06:01,  2.21s/it]\n",
      "Loss: 0.01155:  34%|███▍      | 86/250 [03:10<06:01,  2.21s/it]\n",
      "Loss: 0.01441:  34%|███▍      | 86/250 [03:10<06:01,  2.21s/it]\n",
      "Loss: 0.01328:  34%|███▍      | 86/250 [03:11<06:01,  2.21s/it]\n",
      "Loss: 0.01406:  34%|███▍      | 86/250 [03:11<06:01,  2.21s/it]\n",
      "Loss: 0.01350:  34%|███▍      | 86/250 [03:11<06:01,  2.21s/it]\n",
      "Loss: 0.01496:  34%|███▍      | 86/250 [03:11<06:01,  2.21s/it]\n",
      "Loss: 0.01309:  34%|███▍      | 86/250 [03:11<06:01,  2.21s/it]\n",
      "Loss: 0.01261:  34%|███▍      | 86/250 [03:11<06:01,  2.21s/it]\n",
      "Loss: 0.01273:  34%|███▍      | 86/250 [03:11<06:01,  2.21s/it]\n",
      "Loss: 0.01307:  34%|███▍      | 86/250 [03:11<06:01,  2.21s/it]\n",
      "Loss: 0.01156:  34%|███▍      | 86/250 [03:11<06:01,  2.21s/it]\n",
      "Loss: 0.01232:  34%|███▍      | 86/250 [03:11<06:01,  2.21s/it]\n",
      "Loss: 0.01491:  34%|███▍      | 86/250 [03:12<06:01,  2.21s/it]\n",
      "Loss: 0.01223:  34%|███▍      | 86/250 [03:12<06:01,  2.21s/it]\n",
      "Loss: 0.01430:  34%|███▍      | 86/250 [03:12<06:01,  2.21s/it]\n",
      "Loss: 0.01432:  34%|███▍      | 86/250 [03:12<06:01,  2.21s/it]\n",
      "Loss: 0.01308:  34%|███▍      | 86/250 [03:12<06:01,  2.21s/it]\n",
      "Loss: 0.01366:  34%|███▍      | 86/250 [03:12<06:01,  2.21s/it]\n",
      "Loss: 0.01366:  35%|███▍      | 87/250 [03:12<05:59,  2.21s/it]\n",
      "Loss: 0.01332:  35%|███▍      | 87/250 [03:12<05:59,  2.21s/it]\n",
      "Loss: 0.01318:  35%|███▍      | 87/250 [03:12<05:59,  2.21s/it]\n",
      "Loss: 0.01279:  35%|███▍      | 87/250 [03:12<05:59,  2.21s/it]\n",
      "Loss: 0.01535:  35%|███▍      | 87/250 [03:13<05:59,  2.21s/it]\n",
      "Loss: 0.01422:  35%|███▍      | 87/250 [03:13<05:59,  2.21s/it]\n",
      "Loss: 0.01332:  35%|███▍      | 87/250 [03:13<05:59,  2.21s/it]\n",
      "Loss: 0.01449:  35%|███▍      | 87/250 [03:13<05:59,  2.21s/it]\n",
      "Loss: 0.01296:  35%|███▍      | 87/250 [03:13<05:59,  2.21s/it]\n",
      "Loss: 0.01573:  35%|███▍      | 87/250 [03:13<05:59,  2.21s/it]\n",
      "Loss: 0.01484:  35%|███▍      | 87/250 [03:13<05:59,  2.21s/it]\n",
      "Loss: 0.01343:  35%|███▍      | 87/250 [03:13<05:59,  2.21s/it]\n",
      "Loss: 0.01425:  35%|███▍      | 87/250 [03:13<05:59,  2.21s/it]\n",
      "Loss: 0.01400:  35%|███▍      | 87/250 [03:13<05:59,  2.21s/it]\n",
      "Loss: 0.01353:  35%|███▍      | 87/250 [03:14<05:59,  2.21s/it]\n",
      "Loss: 0.01433:  35%|███▍      | 87/250 [03:14<05:59,  2.21s/it]\n",
      "Loss: 0.01145:  35%|███▍      | 87/250 [03:14<05:59,  2.21s/it]\n",
      "Loss: 0.01353:  35%|███▍      | 87/250 [03:14<05:59,  2.21s/it]\n",
      "Loss: 0.01253:  35%|███▍      | 87/250 [03:14<05:59,  2.21s/it]\n",
      "Loss: 0.01165:  35%|███▍      | 87/250 [03:14<05:59,  2.21s/it]\n",
      "Loss: 0.01271:  35%|███▍      | 87/250 [03:14<05:59,  2.21s/it]\n",
      "Loss: 0.01340:  35%|███▍      | 87/250 [03:14<05:59,  2.21s/it]\n",
      "Loss: 0.01340:  35%|███▌      | 88/250 [03:14<05:55,  2.20s/it]\n",
      "Loss: 0.01267:  35%|███▌      | 88/250 [03:14<05:55,  2.20s/it]\n",
      "Loss: 0.01344:  35%|███▌      | 88/250 [03:14<05:55,  2.20s/it]\n",
      "Loss: 0.01464:  35%|███▌      | 88/250 [03:15<05:55,  2.20s/it]\n",
      "Loss: 0.01170:  35%|███▌      | 88/250 [03:15<05:55,  2.20s/it]\n",
      "Loss: 0.01112:  35%|███▌      | 88/250 [03:15<05:55,  2.20s/it]\n",
      "Loss: 0.01384:  35%|███▌      | 88/250 [03:15<05:55,  2.20s/it]\n",
      "Loss: 0.01074:  35%|███▌      | 88/250 [03:15<05:55,  2.20s/it]\n",
      "Loss: 0.01677:  35%|███▌      | 88/250 [03:15<05:55,  2.20s/it]\n",
      "Loss: 0.01203:  35%|███▌      | 88/250 [03:15<05:55,  2.20s/it]\n",
      "Loss: 0.01487:  35%|███▌      | 88/250 [03:15<05:55,  2.20s/it]\n",
      "Loss: 0.01485:  35%|███▌      | 88/250 [03:15<05:55,  2.20s/it]\n",
      "Loss: 0.01294:  35%|███▌      | 88/250 [03:16<05:55,  2.20s/it]\n",
      "Loss: 0.01395:  35%|███▌      | 88/250 [03:16<05:55,  2.20s/it]\n",
      "Loss: 0.01474:  35%|███▌      | 88/250 [03:16<05:55,  2.20s/it]\n",
      "Loss: 0.01233:  35%|███▌      | 88/250 [03:16<05:55,  2.20s/it]\n",
      "Loss: 0.01508:  35%|███▌      | 88/250 [03:16<05:55,  2.20s/it]\n",
      "Loss: 0.01206:  35%|███▌      | 88/250 [03:16<05:55,  2.20s/it]\n",
      "Loss: 0.01411:  35%|███▌      | 88/250 [03:16<05:55,  2.20s/it]\n",
      "Loss: 0.01422:  35%|███▌      | 88/250 [03:16<05:55,  2.20s/it]\n",
      "Loss: 0.01327:  35%|███▌      | 88/250 [03:16<05:55,  2.20s/it]\n",
      "Loss: 0.01380:  35%|███▌      | 88/250 [03:16<05:55,  2.20s/it]\n",
      "Loss: 0.01380:  36%|███▌      | 89/250 [03:16<05:54,  2.20s/it]\n",
      "Loss: 0.01280:  36%|███▌      | 89/250 [03:17<05:54,  2.20s/it]\n",
      "Loss: 0.01321:  36%|███▌      | 89/250 [03:17<05:54,  2.20s/it]\n",
      "Loss: 0.01485:  36%|███▌      | 89/250 [03:17<05:54,  2.20s/it]\n",
      "Loss: 0.01177:  36%|███▌      | 89/250 [03:17<05:54,  2.20s/it]\n",
      "Loss: 0.01426:  36%|███▌      | 89/250 [03:17<05:54,  2.20s/it]\n",
      "Loss: 0.01349:  36%|███▌      | 89/250 [03:17<05:54,  2.20s/it]\n",
      "Loss: 0.01412:  36%|███▌      | 89/250 [03:17<05:54,  2.20s/it]\n",
      "Loss: 0.01275:  36%|███▌      | 89/250 [03:17<05:54,  2.20s/it]\n",
      "Loss: 0.01507:  36%|███▌      | 89/250 [03:17<05:54,  2.20s/it]\n",
      "Loss: 0.01339:  36%|███▌      | 89/250 [03:18<05:54,  2.20s/it]\n",
      "Loss: 0.01369:  36%|███▌      | 89/250 [03:18<05:54,  2.20s/it]\n",
      "Loss: 0.01400:  36%|███▌      | 89/250 [03:18<05:54,  2.20s/it]\n",
      "Loss: 0.01397:  36%|███▌      | 89/250 [03:18<05:54,  2.20s/it]\n",
      "Loss: 0.01349:  36%|███▌      | 89/250 [03:18<05:54,  2.20s/it]\n",
      "Loss: 0.01330:  36%|███▌      | 89/250 [03:18<05:54,  2.20s/it]\n",
      "Loss: 0.01142:  36%|███▌      | 89/250 [03:18<05:54,  2.20s/it]\n",
      "Loss: 0.01351:  36%|███▌      | 89/250 [03:18<05:54,  2.20s/it]\n",
      "Loss: 0.01286:  36%|███▌      | 89/250 [03:18<05:54,  2.20s/it]\n",
      "Loss: 0.01166:  36%|███▌      | 89/250 [03:19<05:54,  2.20s/it]\n",
      "Loss: 0.01274:  36%|███▌      | 89/250 [03:19<05:54,  2.20s/it]\n",
      "Loss: 0.01325:  36%|███▌      | 89/250 [03:19<05:54,  2.20s/it]\n",
      "Loss: 0.01325:  36%|███▌      | 90/250 [03:19<05:51,  2.20s/it]\n",
      "Loss: 0.01330:  36%|███▌      | 90/250 [03:19<05:51,  2.20s/it]\n",
      "Loss: 0.01275:  36%|███▌      | 90/250 [03:19<05:51,  2.20s/it]\n",
      "Loss: 0.01388:  36%|███▌      | 90/250 [03:19<05:51,  2.20s/it]\n",
      "Loss: 0.01436:  36%|███▌      | 90/250 [03:19<05:51,  2.20s/it]\n",
      "Loss: 0.01421:  36%|███▌      | 90/250 [03:19<05:51,  2.20s/it]\n",
      "Loss: 0.01340:  36%|███▌      | 90/250 [03:19<05:51,  2.20s/it]\n",
      "Loss: 0.01401:  36%|███▌      | 90/250 [03:19<05:51,  2.20s/it]\n",
      "Loss: 0.01238:  36%|███▌      | 90/250 [03:20<05:51,  2.20s/it]\n",
      "Loss: 0.01534:  36%|███▌      | 90/250 [03:20<05:51,  2.20s/it]\n",
      "Loss: 0.01312:  36%|███▌      | 90/250 [03:20<05:51,  2.20s/it]\n",
      "Loss: 0.01206:  36%|███▌      | 90/250 [03:20<05:51,  2.20s/it]\n",
      "Loss: 0.01236:  36%|███▌      | 90/250 [03:20<05:51,  2.20s/it]\n",
      "Loss: 0.01326:  36%|███▌      | 90/250 [03:20<05:51,  2.20s/it]\n",
      "Loss: 0.01177:  36%|███▌      | 90/250 [03:20<05:51,  2.20s/it]\n",
      "Loss: 0.01237:  36%|███▌      | 90/250 [03:20<05:51,  2.20s/it]\n",
      "Loss: 0.01447:  36%|███▌      | 90/250 [03:20<05:51,  2.20s/it]\n",
      "Loss: 0.01223:  36%|███▌      | 90/250 [03:21<05:51,  2.20s/it]\n",
      "Loss: 0.01425:  36%|███▌      | 90/250 [03:21<05:51,  2.20s/it]\n",
      "Loss: 0.01403:  36%|███▌      | 90/250 [03:21<05:51,  2.20s/it]\n",
      "Loss: 0.01328:  36%|███▌      | 90/250 [03:21<05:51,  2.20s/it]\n",
      "Loss: 0.01349:  36%|███▌      | 90/250 [03:21<05:51,  2.20s/it]\n",
      "Loss: 0.01349:  36%|███▋      | 91/250 [03:21<05:51,  2.21s/it]\n",
      "Loss: 0.01252:  36%|███▋      | 91/250 [03:21<05:51,  2.21s/it]\n",
      "Loss: 0.01337:  36%|███▋      | 91/250 [03:21<05:51,  2.21s/it]\n",
      "Loss: 0.01462:  36%|███▋      | 91/250 [03:21<05:51,  2.21s/it]\n",
      "Loss: 0.01167:  36%|███▋      | 91/250 [03:21<05:51,  2.21s/it]\n",
      "Loss: 0.01094:  36%|███▋      | 91/250 [03:21<05:51,  2.21s/it]\n",
      "Loss: 0.01322:  36%|███▋      | 91/250 [03:22<05:51,  2.21s/it]\n",
      "Loss: 0.01464:  36%|███▋      | 91/250 [03:22<05:51,  2.21s/it]\n",
      "Loss: 0.01301:  36%|███▋      | 91/250 [03:22<05:51,  2.21s/it]\n",
      "Loss: 0.01501:  36%|███▋      | 91/250 [03:22<05:51,  2.21s/it]\n",
      "Loss: 0.01320:  36%|███▋      | 91/250 [03:22<05:51,  2.21s/it]\n",
      "Loss: 0.01243:  36%|███▋      | 91/250 [03:22<05:51,  2.21s/it]\n",
      "Loss: 0.01236:  36%|███▋      | 91/250 [03:22<05:51,  2.21s/it]\n",
      "Loss: 0.01287:  36%|███▋      | 91/250 [03:22<05:51,  2.21s/it]\n",
      "Loss: 0.01334:  36%|███▋      | 91/250 [03:22<05:51,  2.21s/it]\n",
      "Loss: 0.01424:  36%|███▋      | 91/250 [03:23<05:51,  2.21s/it]\n",
      "Loss: 0.01280:  36%|███▋      | 91/250 [03:23<05:51,  2.21s/it]\n",
      "Loss: 0.01417:  36%|███▋      | 91/250 [03:23<05:51,  2.21s/it]\n",
      "Loss: 0.01250:  36%|███▋      | 91/250 [03:23<05:51,  2.21s/it]\n",
      "Loss: 0.01314:  36%|███▋      | 91/250 [03:23<05:51,  2.21s/it]\n",
      "Loss: 0.01267:  36%|███▋      | 91/250 [03:23<05:51,  2.21s/it]\n",
      "Loss: 0.01311:  36%|███▋      | 91/250 [03:23<05:51,  2.21s/it]\n",
      "Loss: 0.01311:  37%|███▋      | 92/250 [03:23<05:47,  2.20s/it]\n",
      "Loss: 0.01279:  37%|███▋      | 92/250 [03:23<05:47,  2.20s/it]\n",
      "Loss: 0.01300:  37%|███▋      | 92/250 [03:23<05:47,  2.20s/it]\n",
      "Loss: 0.01475:  37%|███▋      | 92/250 [03:23<05:47,  2.20s/it]\n",
      "Loss: 0.01139:  37%|███▋      | 92/250 [03:24<05:47,  2.20s/it]\n",
      "Loss: 0.01087:  37%|███▋      | 92/250 [03:24<05:47,  2.20s/it]\n",
      "Loss: 0.01364:  37%|███▋      | 92/250 [03:24<05:47,  2.20s/it]\n",
      "Loss: 0.01084:  37%|███▋      | 92/250 [03:24<05:47,  2.20s/it]\n",
      "Loss: 0.01242:  37%|███▋      | 92/250 [03:24<05:47,  2.20s/it]\n",
      "Loss: 0.01502:  37%|███▋      | 92/250 [03:24<05:47,  2.20s/it]\n",
      "Loss: 0.01314:  37%|███▋      | 92/250 [03:24<05:47,  2.20s/it]\n",
      "Loss: 0.01358:  37%|███▋      | 92/250 [03:24<05:47,  2.20s/it]\n",
      "Loss: 0.01411:  37%|███▋      | 92/250 [03:24<05:47,  2.20s/it]\n",
      "Loss: 0.01414:  37%|███▋      | 92/250 [03:24<05:47,  2.20s/it]\n",
      "Loss: 0.01321:  37%|███▋      | 92/250 [03:25<05:47,  2.20s/it]\n",
      "Loss: 0.01400:  37%|███▋      | 92/250 [03:25<05:47,  2.20s/it]\n",
      "Loss: 0.01218:  37%|███▋      | 92/250 [03:25<05:47,  2.20s/it]\n",
      "Loss: 0.01321:  37%|███▋      | 92/250 [03:25<05:47,  2.20s/it]\n",
      "Loss: 0.01279:  37%|███▋      | 92/250 [03:25<05:47,  2.20s/it]\n",
      "Loss: 0.01176:  37%|███▋      | 92/250 [03:25<05:47,  2.20s/it]\n",
      "Loss: 0.01278:  37%|███▋      | 92/250 [03:25<05:47,  2.20s/it]\n",
      "Loss: 0.01331:  37%|███▋      | 92/250 [03:25<05:47,  2.20s/it]\n",
      "Loss: 0.01331:  37%|███▋      | 93/250 [03:25<05:45,  2.20s/it]\n",
      "Loss: 0.01249:  37%|███▋      | 93/250 [03:25<05:45,  2.20s/it]\n",
      "Loss: 0.01336:  37%|███▋      | 93/250 [03:25<05:45,  2.20s/it]\n",
      "Loss: 0.01441:  37%|███▋      | 93/250 [03:26<05:45,  2.20s/it]\n",
      "Loss: 0.01178:  37%|███▋      | 93/250 [03:26<05:45,  2.20s/it]\n",
      "Loss: 0.01110:  37%|███▋      | 93/250 [03:26<05:45,  2.20s/it]\n",
      "Loss: 0.01366:  37%|███▋      | 93/250 [03:26<05:45,  2.20s/it]\n",
      "Loss: 0.01060:  37%|███▋      | 93/250 [03:26<05:45,  2.20s/it]\n",
      "Loss: 0.01257:  37%|███▋      | 93/250 [03:26<05:45,  2.20s/it]\n",
      "Loss: 0.01168:  37%|███▋      | 93/250 [03:26<05:45,  2.20s/it]\n",
      "Loss: 0.01199:  37%|███▋      | 93/250 [03:26<05:45,  2.20s/it]\n",
      "Loss: 0.01372:  37%|███▋      | 93/250 [03:26<05:45,  2.20s/it]\n",
      "Loss: 0.01259:  37%|███▋      | 93/250 [03:27<05:45,  2.20s/it]\n",
      "Loss: 0.01190:  37%|███▋      | 93/250 [03:27<05:45,  2.20s/it]\n",
      "Loss: 0.01156:  37%|███▋      | 93/250 [03:27<05:45,  2.20s/it]\n",
      "Loss: 0.01344:  37%|███▋      | 93/250 [03:27<05:45,  2.20s/it]\n",
      "Loss: 0.01098:  37%|███▋      | 93/250 [03:27<05:45,  2.20s/it]\n",
      "Loss: 0.01348:  37%|███▋      | 93/250 [03:27<05:45,  2.20s/it]\n",
      "Loss: 0.01233:  37%|███▋      | 93/250 [03:27<05:45,  2.20s/it]\n",
      "Loss: 0.01189:  37%|███▋      | 93/250 [03:27<05:45,  2.20s/it]\n",
      "Loss: 0.01250:  37%|███▋      | 93/250 [03:27<05:45,  2.20s/it]\n",
      "Loss: 0.01361:  37%|███▋      | 93/250 [03:27<05:45,  2.20s/it]\n",
      "Loss: 0.01361:  38%|███▊      | 94/250 [03:27<05:43,  2.20s/it]\n",
      "Loss: 0.01237:  38%|███▊      | 94/250 [03:28<05:43,  2.20s/it]\n",
      "Loss: 0.01276:  38%|███▊      | 94/250 [03:28<05:43,  2.20s/it]\n",
      "Loss: 0.01441:  38%|███▊      | 94/250 [03:28<05:43,  2.20s/it]\n",
      "Loss: 0.01150:  38%|███▊      | 94/250 [03:28<05:43,  2.20s/it]\n",
      "Loss: 0.01060:  38%|███▊      | 94/250 [03:28<05:43,  2.20s/it]\n",
      "Loss: 0.01317:  38%|███▊      | 94/250 [03:28<05:43,  2.20s/it]\n",
      "Loss: 0.01407:  38%|███▊      | 94/250 [03:28<05:43,  2.20s/it]\n",
      "Loss: 0.01242:  38%|███▊      | 94/250 [03:28<05:43,  2.20s/it]\n",
      "Loss: 0.01480:  38%|███▊      | 94/250 [03:28<05:43,  2.20s/it]\n",
      "Loss: 0.01355:  38%|███▊      | 94/250 [03:29<05:43,  2.20s/it]\n",
      "Loss: 0.01204:  38%|███▊      | 94/250 [03:29<05:43,  2.20s/it]\n",
      "Loss: 0.01208:  38%|███▊      | 94/250 [03:29<05:43,  2.20s/it]\n",
      "Loss: 0.01250:  38%|███▊      | 94/250 [03:29<05:43,  2.20s/it]\n",
      "Loss: 0.01140:  38%|███▊      | 94/250 [03:29<05:43,  2.20s/it]\n",
      "Loss: 0.01215:  38%|███▊      | 94/250 [03:29<05:43,  2.20s/it]\n",
      "Loss: 0.01460:  38%|███▊      | 94/250 [03:29<05:43,  2.20s/it]\n",
      "Loss: 0.01210:  38%|███▊      | 94/250 [03:29<05:43,  2.20s/it]\n",
      "Loss: 0.01406:  38%|███▊      | 94/250 [03:29<05:43,  2.20s/it]\n",
      "Loss: 0.01367:  38%|███▊      | 94/250 [03:29<05:43,  2.20s/it]\n",
      "Loss: 0.01315:  38%|███▊      | 94/250 [03:30<05:43,  2.20s/it]\n",
      "Loss: 0.01320:  38%|███▊      | 94/250 [03:30<05:43,  2.20s/it]\n",
      "Loss: 0.01320:  38%|███▊      | 95/250 [03:30<05:41,  2.20s/it]\n",
      "Loss: 0.01248:  38%|███▊      | 95/250 [03:30<05:41,  2.20s/it]\n",
      "Loss: 0.01284:  38%|███▊      | 95/250 [03:30<05:41,  2.20s/it]\n",
      "Loss: 0.01444:  38%|███▊      | 95/250 [03:30<05:41,  2.20s/it]\n",
      "Loss: 0.01140:  38%|███▊      | 95/250 [03:30<05:41,  2.20s/it]\n",
      "Loss: 0.01092:  38%|███▊      | 95/250 [03:30<05:41,  2.20s/it]\n",
      "Loss: 0.01352:  38%|███▊      | 95/250 [03:30<05:41,  2.20s/it]\n",
      "Loss: 0.01066:  38%|███▊      | 95/250 [03:30<05:41,  2.20s/it]\n",
      "Loss: 0.01649:  38%|███▊      | 95/250 [03:31<05:41,  2.20s/it]\n",
      "Loss: 0.01494:  38%|███▊      | 95/250 [03:31<05:41,  2.20s/it]\n",
      "Loss: 0.01330:  38%|███▊      | 95/250 [03:31<05:41,  2.20s/it]\n",
      "Loss: 0.01215:  38%|███▊      | 95/250 [03:31<05:41,  2.20s/it]\n",
      "Loss: 0.01219:  38%|███▊      | 95/250 [03:31<05:41,  2.20s/it]\n",
      "Loss: 0.01258:  38%|███▊      | 95/250 [03:31<05:41,  2.20s/it]\n",
      "Loss: 0.01155:  38%|███▊      | 95/250 [03:31<05:41,  2.20s/it]\n",
      "Loss: 0.01391:  38%|███▊      | 95/250 [03:31<05:41,  2.20s/it]\n",
      "Loss: 0.01216:  38%|███▊      | 95/250 [03:31<05:41,  2.20s/it]\n",
      "Loss: 0.01369:  38%|███▊      | 95/250 [03:31<05:41,  2.20s/it]\n",
      "Loss: 0.01275:  38%|███▊      | 95/250 [03:32<05:41,  2.20s/it]\n",
      "Loss: 0.01338:  38%|███▊      | 95/250 [03:32<05:41,  2.20s/it]\n",
      "Loss: 0.01275:  38%|███▊      | 95/250 [03:32<05:41,  2.20s/it]\n",
      "Loss: 0.01340:  38%|███▊      | 95/250 [03:32<05:41,  2.20s/it]\n",
      "Loss: 0.01340:  38%|███▊      | 96/250 [03:32<05:37,  2.19s/it]\n",
      "Loss: 0.01229:  38%|███▊      | 96/250 [03:32<05:37,  2.19s/it]\n",
      "Loss: 0.01285:  38%|███▊      | 96/250 [03:32<05:37,  2.19s/it]\n",
      "Loss: 0.01447:  38%|███▊      | 96/250 [03:32<05:37,  2.19s/it]\n",
      "Loss: 0.01129:  38%|███▊      | 96/250 [03:32<05:37,  2.19s/it]\n",
      "Loss: 0.01110:  38%|███▊      | 96/250 [03:32<05:37,  2.19s/it]\n",
      "Loss: 0.01330:  38%|███▊      | 96/250 [03:32<05:37,  2.19s/it]\n",
      "Loss: 0.01033:  38%|███▊      | 96/250 [03:33<05:37,  2.19s/it]\n",
      "Loss: 0.01655:  38%|███▊      | 96/250 [03:33<05:37,  2.19s/it]\n",
      "Loss: 0.01121:  38%|███▊      | 96/250 [03:33<05:37,  2.19s/it]\n",
      "Loss: 0.01453:  38%|███▊      | 96/250 [03:33<05:37,  2.19s/it]\n",
      "Loss: 0.01478:  38%|███▊      | 96/250 [03:33<05:37,  2.19s/it]\n",
      "Loss: 0.01275:  38%|███▊      | 96/250 [03:33<05:37,  2.19s/it]\n",
      "Loss: 0.01394:  38%|███▊      | 96/250 [03:33<05:37,  2.19s/it]\n",
      "Loss: 0.01408:  38%|███▊      | 96/250 [03:33<05:37,  2.19s/it]\n",
      "Loss: 0.01206:  38%|███▊      | 96/250 [03:33<05:37,  2.19s/it]\n",
      "Loss: 0.01154:  38%|███▊      | 96/250 [03:34<05:37,  2.19s/it]\n",
      "Loss: 0.01186:  38%|███▊      | 96/250 [03:34<05:37,  2.19s/it]\n",
      "Loss: 0.01421:  38%|███▊      | 96/250 [03:34<05:37,  2.19s/it]\n",
      "Loss: 0.01364:  38%|███▊      | 96/250 [03:34<05:37,  2.19s/it]\n",
      "Loss: 0.01279:  38%|███▊      | 96/250 [03:34<05:37,  2.19s/it]\n",
      "Loss: 0.01377:  38%|███▊      | 96/250 [03:34<05:37,  2.19s/it]\n",
      "Loss: 0.01377:  39%|███▉      | 97/250 [03:34<05:36,  2.20s/it]\n",
      "Loss: 0.01268:  39%|███▉      | 97/250 [03:34<05:36,  2.20s/it]\n",
      "Loss: 0.01324:  39%|███▉      | 97/250 [03:34<05:36,  2.20s/it]\n",
      "Loss: 0.01456:  39%|███▉      | 97/250 [03:34<05:36,  2.20s/it]\n",
      "Loss: 0.01137:  39%|███▉      | 97/250 [03:35<05:36,  2.20s/it]\n",
      "Loss: 0.01114:  39%|███▉      | 97/250 [03:35<05:36,  2.20s/it]\n",
      "Loss: 0.01318:  39%|███▉      | 97/250 [03:35<05:36,  2.20s/it]\n",
      "Loss: 0.01052:  39%|███▉      | 97/250 [03:35<05:36,  2.20s/it]\n",
      "Loss: 0.01637:  39%|███▉      | 97/250 [03:35<05:36,  2.20s/it]\n",
      "Loss: 0.01150:  39%|███▉      | 97/250 [03:35<05:36,  2.20s/it]\n",
      "Loss: 0.01418:  39%|███▉      | 97/250 [03:35<05:36,  2.20s/it]\n",
      "Loss: 0.01510:  39%|███▉      | 97/250 [03:35<05:36,  2.20s/it]\n",
      "Loss: 0.01283:  39%|███▉      | 97/250 [03:35<05:36,  2.20s/it]\n",
      "Loss: 0.01328:  39%|███▉      | 97/250 [03:35<05:36,  2.20s/it]\n",
      "Loss: 0.01474:  39%|███▉      | 97/250 [03:36<05:36,  2.20s/it]\n",
      "Loss: 0.01210:  39%|███▉      | 97/250 [03:36<05:36,  2.20s/it]\n",
      "Loss: 0.01420:  39%|███▉      | 97/250 [03:36<05:36,  2.20s/it]\n",
      "Loss: 0.01194:  39%|███▉      | 97/250 [03:36<05:36,  2.20s/it]\n",
      "Loss: 0.01418:  39%|███▉      | 97/250 [03:36<05:36,  2.20s/it]\n",
      "Loss: 0.01364:  39%|███▉      | 97/250 [03:36<05:36,  2.20s/it]\n",
      "Loss: 0.01308:  39%|███▉      | 97/250 [03:36<05:36,  2.20s/it]\n",
      "Loss: 0.01342:  39%|███▉      | 97/250 [03:36<05:36,  2.20s/it]\n",
      "Loss: 0.01342:  39%|███▉      | 98/250 [03:36<05:35,  2.21s/it]\n",
      "Loss: 0.01289:  39%|███▉      | 98/250 [03:36<05:35,  2.21s/it]\n",
      "Loss: 0.01283:  39%|███▉      | 98/250 [03:37<05:35,  2.21s/it]\n",
      "Loss: 0.01447:  39%|███▉      | 98/250 [03:37<05:35,  2.21s/it]\n",
      "Loss: 0.01104:  39%|███▉      | 98/250 [03:37<05:35,  2.21s/it]\n",
      "Loss: 0.01051:  39%|███▉      | 98/250 [03:37<05:35,  2.21s/it]\n",
      "Loss: 0.01310:  39%|███▉      | 98/250 [03:37<05:35,  2.21s/it]\n",
      "Loss: 0.01060:  39%|███▉      | 98/250 [03:37<05:35,  2.21s/it]\n",
      "Loss: 0.01631:  39%|███▉      | 98/250 [03:37<05:35,  2.21s/it]\n",
      "Loss: 0.01116:  39%|███▉      | 98/250 [03:37<05:35,  2.21s/it]\n",
      "Loss: 0.01450:  39%|███▉      | 98/250 [03:37<05:35,  2.21s/it]\n",
      "Loss: 0.01573:  39%|███▉      | 98/250 [03:37<05:35,  2.21s/it]\n",
      "Loss: 0.01267:  39%|███▉      | 98/250 [03:38<05:35,  2.21s/it]\n",
      "Loss: 0.01379:  39%|███▉      | 98/250 [03:38<05:35,  2.21s/it]\n",
      "Loss: 0.01441:  39%|███▉      | 98/250 [03:38<05:35,  2.21s/it]\n",
      "Loss: 0.01223:  39%|███▉      | 98/250 [03:38<05:35,  2.21s/it]\n",
      "Loss: 0.01139:  39%|███▉      | 98/250 [03:38<05:35,  2.21s/it]\n",
      "Loss: 0.01236:  39%|███▉      | 98/250 [03:38<05:35,  2.21s/it]\n",
      "Loss: 0.01390:  39%|███▉      | 98/250 [03:38<05:35,  2.21s/it]\n",
      "Loss: 0.01371:  39%|███▉      | 98/250 [03:38<05:35,  2.21s/it]\n",
      "Loss: 0.01303:  39%|███▉      | 98/250 [03:38<05:35,  2.21s/it]\n",
      "Loss: 0.01299:  39%|███▉      | 98/250 [03:39<05:35,  2.21s/it]\n",
      "Loss: 0.01299:  40%|███▉      | 99/250 [03:39<05:33,  2.21s/it]\n",
      "Loss: 0.01356:  40%|███▉      | 99/250 [03:39<05:33,  2.21s/it]\n",
      "Loss: 0.01261:  40%|███▉      | 99/250 [03:39<05:33,  2.21s/it]\n",
      "Loss: 0.01263:  40%|███▉      | 99/250 [03:39<05:33,  2.21s/it]\n",
      "Loss: 0.01429:  40%|███▉      | 99/250 [03:39<05:33,  2.21s/it]\n",
      "Loss: 0.01407:  40%|███▉      | 99/250 [03:39<05:33,  2.21s/it]\n",
      "Loss: 0.01299:  40%|███▉      | 99/250 [03:39<05:33,  2.21s/it]\n",
      "Loss: 0.01383:  40%|███▉      | 99/250 [03:39<05:33,  2.21s/it]\n",
      "Loss: 0.01327:  40%|███▉      | 99/250 [03:39<05:33,  2.21s/it]\n",
      "Loss: 0.01185:  40%|███▉      | 99/250 [03:40<05:33,  2.21s/it]\n",
      "Loss: 0.01425:  40%|███▉      | 99/250 [03:40<05:33,  2.21s/it]\n",
      "Loss: 0.01300:  40%|███▉      | 99/250 [03:40<05:33,  2.21s/it]\n",
      "Loss: 0.01232:  40%|███▉      | 99/250 [03:40<05:33,  2.21s/it]\n",
      "Loss: 0.01214:  40%|███▉      | 99/250 [03:40<05:33,  2.21s/it]\n",
      "Loss: 0.01077:  40%|███▉      | 99/250 [03:40<05:33,  2.21s/it]\n",
      "Loss: 0.01275:  40%|███▉      | 99/250 [03:40<05:33,  2.21s/it]\n",
      "Loss: 0.01233:  40%|███▉      | 99/250 [03:40<05:33,  2.21s/it]\n",
      "Loss: 0.01252:  40%|███▉      | 99/250 [03:40<05:33,  2.21s/it]\n",
      "Loss: 0.01652:  40%|███▉      | 99/250 [03:40<05:33,  2.21s/it]\n",
      "Loss: 0.01227:  40%|███▉      | 99/250 [03:41<05:33,  2.21s/it]\n",
      "Loss: 0.01305:  40%|███▉      | 99/250 [03:41<05:33,  2.21s/it]\n",
      "Loss: 0.01305:  40%|████      | 100/250 [03:41<05:29,  2.20s/it]\n",
      "Loss: 0.01276:  40%|████      | 100/250 [03:41<05:29,  2.20s/it]\n",
      "Loss: 0.01280:  40%|████      | 100/250 [03:41<05:29,  2.20s/it]\n",
      "Loss: 0.01252:  40%|████      | 100/250 [03:41<05:29,  2.20s/it]\n",
      "Loss: 0.01431:  40%|████      | 100/250 [03:41<05:29,  2.20s/it]\n",
      "Loss: 0.01387:  40%|████      | 100/250 [03:41<05:29,  2.20s/it]\n",
      "Loss: 0.01289:  40%|████      | 100/250 [03:41<05:29,  2.20s/it]\n",
      "Loss: 0.01385:  40%|████      | 100/250 [03:41<05:29,  2.20s/it]\n",
      "Loss: 0.01214:  40%|████      | 100/250 [03:42<05:29,  2.20s/it]\n",
      "Loss: 0.01480:  40%|████      | 100/250 [03:42<05:29,  2.20s/it]\n",
      "Loss: 0.01300:  40%|████      | 100/250 [03:42<05:29,  2.20s/it]\n",
      "Loss: 0.01202:  40%|████      | 100/250 [03:42<05:29,  2.20s/it]\n",
      "Loss: 0.01191:  40%|████      | 100/250 [03:42<05:29,  2.20s/it]\n",
      "Loss: 0.01252:  40%|████      | 100/250 [03:42<05:29,  2.20s/it]\n",
      "Loss: 0.01115:  40%|████      | 100/250 [03:42<05:29,  2.20s/it]\n",
      "Loss: 0.01152:  40%|████      | 100/250 [03:42<05:29,  2.20s/it]\n",
      "Loss: 0.01447:  40%|████      | 100/250 [03:42<05:29,  2.20s/it]\n",
      "Loss: 0.01198:  40%|████      | 100/250 [03:43<05:29,  2.20s/it]\n",
      "Loss: 0.01388:  40%|████      | 100/250 [03:43<05:29,  2.20s/it]\n",
      "Loss: 0.01414:  40%|████      | 100/250 [03:43<05:29,  2.20s/it]\n",
      "Loss: 0.01243:  40%|████      | 100/250 [03:43<05:29,  2.20s/it]\n",
      "Loss: 0.01328:  40%|████      | 100/250 [03:43<05:29,  2.20s/it]\n",
      "Loss: 0.01328:  40%|████      | 101/250 [03:43<05:31,  2.23s/it]\n",
      "Loss: 0.01335:  40%|████      | 101/250 [03:43<05:31,  2.23s/it]\n",
      "Loss: 0.01242:  40%|████      | 101/250 [03:43<05:31,  2.23s/it]\n",
      "Loss: 0.01242:  40%|████      | 101/250 [03:43<05:31,  2.23s/it]\n",
      "Loss: 0.01444:  40%|████      | 101/250 [03:43<05:31,  2.23s/it]\n",
      "Loss: 0.01426:  40%|████      | 101/250 [03:44<05:31,  2.23s/it]\n",
      "Loss: 0.01285:  40%|████      | 101/250 [03:44<05:31,  2.23s/it]\n",
      "Loss: 0.01410:  40%|████      | 101/250 [03:44<05:31,  2.23s/it]\n",
      "Loss: 0.01224:  40%|████      | 101/250 [03:44<05:31,  2.23s/it]\n",
      "Loss: 0.01493:  40%|████      | 101/250 [03:44<05:31,  2.23s/it]\n",
      "Loss: 0.01300:  40%|████      | 101/250 [03:44<05:31,  2.23s/it]\n",
      "Loss: 0.01217:  40%|████      | 101/250 [03:44<05:31,  2.23s/it]\n",
      "Loss: 0.01218:  40%|████      | 101/250 [03:44<05:31,  2.23s/it]\n",
      "Loss: 0.01253:  40%|████      | 101/250 [03:44<05:31,  2.23s/it]\n",
      "Loss: 0.01118:  40%|████      | 101/250 [03:44<05:31,  2.23s/it]\n",
      "Loss: 0.01208:  40%|████      | 101/250 [03:45<05:31,  2.23s/it]\n",
      "Loss: 0.01410:  40%|████      | 101/250 [03:45<05:31,  2.23s/it]\n",
      "Loss: 0.01142:  40%|████      | 101/250 [03:45<05:31,  2.23s/it]\n",
      "Loss: 0.01363:  40%|████      | 101/250 [03:45<05:31,  2.23s/it]\n",
      "Loss: 0.01348:  40%|████      | 101/250 [03:45<05:31,  2.23s/it]\n",
      "Loss: 0.01266:  40%|████      | 101/250 [03:45<05:31,  2.23s/it]\n",
      "Loss: 0.01298:  40%|████      | 101/250 [03:45<05:31,  2.23s/it]\n",
      "Loss: 0.01298:  41%|████      | 102/250 [03:45<05:28,  2.22s/it]\n",
      "Loss: 0.01211:  41%|████      | 102/250 [03:45<05:28,  2.22s/it]\n",
      "Loss: 0.01262:  41%|████      | 102/250 [03:45<05:28,  2.22s/it]\n",
      "Loss: 0.01409:  41%|████      | 102/250 [03:46<05:28,  2.22s/it]\n",
      "Loss: 0.01137:  41%|████      | 102/250 [03:46<05:28,  2.22s/it]\n",
      "Loss: 0.01151:  41%|████      | 102/250 [03:46<05:28,  2.22s/it]\n",
      "Loss: 0.01312:  41%|████      | 102/250 [03:46<05:28,  2.22s/it]\n",
      "Loss: 0.01372:  41%|████      | 102/250 [03:46<05:28,  2.22s/it]\n",
      "Loss: 0.01238:  41%|████      | 102/250 [03:46<05:28,  2.22s/it]\n",
      "Loss: 0.01494:  41%|████      | 102/250 [03:46<05:28,  2.22s/it]\n",
      "Loss: 0.01287:  41%|████      | 102/250 [03:46<05:28,  2.22s/it]\n",
      "Loss: 0.01194:  41%|████      | 102/250 [03:46<05:28,  2.22s/it]\n",
      "Loss: 0.01162:  41%|████      | 102/250 [03:46<05:28,  2.22s/it]\n",
      "Loss: 0.01242:  41%|████      | 102/250 [03:47<05:28,  2.22s/it]\n",
      "Loss: 0.01141:  41%|████      | 102/250 [03:47<05:28,  2.22s/it]\n",
      "Loss: 0.01181:  41%|████      | 102/250 [03:47<05:28,  2.22s/it]\n",
      "Loss: 0.01426:  41%|████      | 102/250 [03:47<05:28,  2.22s/it]\n",
      "Loss: 0.01158:  41%|████      | 102/250 [03:47<05:28,  2.22s/it]\n",
      "Loss: 0.01398:  41%|████      | 102/250 [03:47<05:28,  2.22s/it]\n",
      "Loss: 0.01345:  41%|████      | 102/250 [03:47<05:28,  2.22s/it]\n",
      "Loss: 0.01270:  41%|████      | 102/250 [03:47<05:28,  2.22s/it]\n",
      "Loss: 0.01283:  41%|████      | 102/250 [03:47<05:28,  2.22s/it]\n",
      "Loss: 0.01283:  41%|████      | 103/250 [03:47<05:26,  2.22s/it]\n",
      "Loss: 0.01190:  41%|████      | 103/250 [03:48<05:26,  2.22s/it]\n",
      "Loss: 0.01312:  41%|████      | 103/250 [03:48<05:26,  2.22s/it]\n",
      "Loss: 0.01441:  41%|████      | 103/250 [03:48<05:26,  2.22s/it]\n",
      "Loss: 0.01127:  41%|████      | 103/250 [03:48<05:26,  2.22s/it]\n",
      "Loss: 0.01406:  41%|████      | 103/250 [03:48<05:26,  2.22s/it]\n",
      "Loss: 0.01280:  41%|████      | 103/250 [03:48<05:26,  2.22s/it]\n",
      "Loss: 0.01391:  41%|████      | 103/250 [03:48<05:26,  2.22s/it]\n",
      "Loss: 0.01211:  41%|████      | 103/250 [03:48<05:26,  2.22s/it]\n",
      "Loss: 0.01476:  41%|████      | 103/250 [03:48<05:26,  2.22s/it]\n",
      "Loss: 0.01332:  41%|████      | 103/250 [03:48<05:26,  2.22s/it]\n",
      "Loss: 0.01244:  41%|████      | 103/250 [03:49<05:26,  2.22s/it]\n",
      "Loss: 0.01364:  41%|████      | 103/250 [03:49<05:26,  2.22s/it]\n",
      "Loss: 0.01409:  41%|████      | 103/250 [03:49<05:26,  2.22s/it]\n",
      "Loss: 0.01124:  41%|████      | 103/250 [03:49<05:26,  2.22s/it]\n",
      "Loss: 0.01333:  41%|████      | 103/250 [03:49<05:26,  2.22s/it]\n",
      "Loss: 0.01076:  41%|████      | 103/250 [03:49<05:26,  2.22s/it]\n",
      "Loss: 0.01409:  41%|████      | 103/250 [03:49<05:26,  2.22s/it]\n",
      "Loss: 0.01230:  41%|████      | 103/250 [03:49<05:26,  2.22s/it]\n",
      "Loss: 0.01111:  41%|████      | 103/250 [03:49<05:26,  2.22s/it]\n",
      "Loss: 0.01265:  41%|████      | 103/250 [03:50<05:26,  2.22s/it]\n",
      "Loss: 0.01290:  41%|████      | 103/250 [03:50<05:26,  2.22s/it]\n",
      "Loss: 0.01290:  42%|████▏     | 104/250 [03:50<05:23,  2.21s/it]\n",
      "Loss: 0.01189:  42%|████▏     | 104/250 [03:50<05:23,  2.21s/it]\n",
      "Loss: 0.01278:  42%|████▏     | 104/250 [03:50<05:23,  2.21s/it]\n",
      "Loss: 0.01464:  42%|████▏     | 104/250 [03:50<05:23,  2.21s/it]\n",
      "Loss: 0.01115:  42%|████▏     | 104/250 [03:50<05:23,  2.21s/it]\n",
      "Loss: 0.01083:  42%|████▏     | 104/250 [03:50<05:23,  2.21s/it]\n",
      "Loss: 0.01301:  42%|████▏     | 104/250 [03:50<05:23,  2.21s/it]\n",
      "Loss: 0.01349:  42%|████▏     | 104/250 [03:50<05:23,  2.21s/it]\n",
      "Loss: 0.01199:  42%|████▏     | 104/250 [03:50<05:23,  2.21s/it]\n",
      "Loss: 0.01478:  42%|████▏     | 104/250 [03:51<05:23,  2.21s/it]\n",
      "Loss: 0.01300:  42%|████▏     | 104/250 [03:51<05:23,  2.21s/it]\n",
      "Loss: 0.01177:  42%|████▏     | 104/250 [03:51<05:23,  2.21s/it]\n",
      "Loss: 0.01223:  42%|████▏     | 104/250 [03:51<05:23,  2.21s/it]\n",
      "Loss: 0.01236:  42%|████▏     | 104/250 [03:51<05:23,  2.21s/it]\n",
      "Loss: 0.01116:  42%|████▏     | 104/250 [03:51<05:23,  2.21s/it]\n",
      "Loss: 0.01184:  42%|████▏     | 104/250 [03:51<05:23,  2.21s/it]\n",
      "Loss: 0.01387:  42%|████▏     | 104/250 [03:51<05:23,  2.21s/it]\n",
      "Loss: 0.01203:  42%|████▏     | 104/250 [03:51<05:23,  2.21s/it]\n",
      "Loss: 0.01380:  42%|████▏     | 104/250 [03:52<05:23,  2.21s/it]\n",
      "Loss: 0.01365:  42%|████▏     | 104/250 [03:52<05:23,  2.21s/it]\n",
      "Loss: 0.01259:  42%|████▏     | 104/250 [03:52<05:23,  2.21s/it]\n",
      "Loss: 0.01273:  42%|████▏     | 104/250 [03:52<05:23,  2.21s/it]\n",
      "Loss: 0.01273:  42%|████▏     | 105/250 [03:52<05:21,  2.21s/it]\n",
      "Loss: 0.01193:  42%|████▏     | 105/250 [03:52<05:21,  2.21s/it]\n",
      "Loss: 0.01293:  42%|████▏     | 105/250 [03:52<05:21,  2.21s/it]\n",
      "Loss: 0.01460:  42%|████▏     | 105/250 [03:52<05:21,  2.21s/it]\n",
      "Loss: 0.01123:  42%|████▏     | 105/250 [03:52<05:21,  2.21s/it]\n",
      "Loss: 0.01091:  42%|████▏     | 105/250 [03:52<05:21,  2.21s/it]\n",
      "Loss: 0.01313:  42%|████▏     | 105/250 [03:52<05:21,  2.21s/it]\n",
      "Loss: 0.01064:  42%|████▏     | 105/250 [03:53<05:21,  2.21s/it]\n",
      "Loss: 0.01587:  42%|████▏     | 105/250 [03:53<05:21,  2.21s/it]\n",
      "Loss: 0.01084:  42%|████▏     | 105/250 [03:53<05:21,  2.21s/it]\n",
      "Loss: 0.01433:  42%|████▏     | 105/250 [03:53<05:21,  2.21s/it]\n",
      "Loss: 0.01414:  42%|████▏     | 105/250 [03:53<05:21,  2.21s/it]\n",
      "Loss: 0.01270:  42%|████▏     | 105/250 [03:53<05:21,  2.21s/it]\n",
      "Loss: 0.01335:  42%|████▏     | 105/250 [03:53<05:21,  2.21s/it]\n",
      "Loss: 0.01427:  42%|████▏     | 105/250 [03:53<05:21,  2.21s/it]\n",
      "Loss: 0.01197:  42%|████▏     | 105/250 [03:53<05:21,  2.21s/it]\n",
      "Loss: 0.01126:  42%|████▏     | 105/250 [03:53<05:21,  2.21s/it]\n",
      "Loss: 0.01251:  42%|████▏     | 105/250 [03:54<05:21,  2.21s/it]\n",
      "Loss: 0.01265:  42%|████▏     | 105/250 [03:54<05:21,  2.21s/it]\n",
      "Loss: 0.01573:  42%|████▏     | 105/250 [03:54<05:21,  2.21s/it]\n",
      "Loss: 0.01182:  42%|████▏     | 105/250 [03:54<05:21,  2.21s/it]\n",
      "Loss: 0.01262:  42%|████▏     | 105/250 [03:54<05:21,  2.21s/it]\n",
      "Loss: 0.01262:  42%|████▏     | 106/250 [03:54<05:19,  2.22s/it]\n",
      "Loss: 0.01318:  42%|████▏     | 106/250 [03:54<05:19,  2.22s/it]\n",
      "Loss: 0.01225:  42%|████▏     | 106/250 [03:54<05:19,  2.22s/it]\n",
      "Loss: 0.01247:  42%|████▏     | 106/250 [03:54<05:19,  2.22s/it]\n",
      "Loss: 0.01398:  42%|████▏     | 106/250 [03:54<05:19,  2.22s/it]\n",
      "Loss: 0.01008:  42%|████▏     | 106/250 [03:55<05:19,  2.22s/it]\n",
      "Loss: 0.01114:  42%|████▏     | 106/250 [03:55<05:19,  2.22s/it]\n",
      "Loss: 0.01135:  42%|████▏     | 106/250 [03:55<05:19,  2.22s/it]\n",
      "Loss: 0.01293:  42%|████▏     | 106/250 [03:55<05:19,  2.22s/it]\n",
      "Loss: 0.01173:  42%|████▏     | 106/250 [03:55<05:19,  2.22s/it]\n",
      "Loss: 0.01414:  42%|████▏     | 106/250 [03:55<05:19,  2.22s/it]\n",
      "Loss: 0.01285:  42%|████▏     | 106/250 [03:55<05:19,  2.22s/it]\n",
      "Loss: 0.01318:  42%|████▏     | 106/250 [03:55<05:19,  2.22s/it]\n",
      "Loss: 0.01206:  42%|████▏     | 106/250 [03:55<05:19,  2.22s/it]\n",
      "Loss: 0.01099:  42%|████▏     | 106/250 [03:56<05:19,  2.22s/it]\n",
      "Loss: 0.01301:  42%|████▏     | 106/250 [03:56<05:19,  2.22s/it]\n",
      "Loss: 0.01092:  42%|████▏     | 106/250 [03:56<05:19,  2.22s/it]\n",
      "Loss: 0.01300:  42%|████▏     | 106/250 [03:56<05:19,  2.22s/it]\n",
      "Loss: 0.01226:  42%|████▏     | 106/250 [03:56<05:19,  2.22s/it]\n",
      "Loss: 0.01133:  42%|████▏     | 106/250 [03:56<05:19,  2.22s/it]\n",
      "Loss: 0.01238:  42%|████▏     | 106/250 [03:56<05:19,  2.22s/it]\n",
      "Loss: 0.01294:  42%|████▏     | 106/250 [03:56<05:19,  2.22s/it]\n",
      "Loss: 0.01294:  43%|████▎     | 107/250 [03:56<05:16,  2.21s/it]\n",
      "Loss: 0.01207:  43%|████▎     | 107/250 [03:56<05:16,  2.21s/it]\n",
      "Loss: 0.01251:  43%|████▎     | 107/250 [03:56<05:16,  2.21s/it]\n",
      "Loss: 0.01410:  43%|████▎     | 107/250 [03:57<05:16,  2.21s/it]\n",
      "Loss: 0.01075:  43%|████▎     | 107/250 [03:57<05:16,  2.21s/it]\n",
      "Loss: 0.01066:  43%|████▎     | 107/250 [03:57<05:16,  2.21s/it]\n",
      "Loss: 0.01288:  43%|████▎     | 107/250 [03:57<05:16,  2.21s/it]\n",
      "Loss: 0.01045:  43%|████▎     | 107/250 [03:57<05:16,  2.21s/it]\n",
      "Loss: 0.01592:  43%|████▎     | 107/250 [03:57<05:16,  2.21s/it]\n",
      "Loss: 0.01094:  43%|████▎     | 107/250 [03:57<05:16,  2.21s/it]\n",
      "Loss: 0.01462:  43%|████▎     | 107/250 [03:57<05:16,  2.21s/it]\n",
      "Loss: 0.01459:  43%|████▎     | 107/250 [03:57<05:16,  2.21s/it]\n",
      "Loss: 0.01234:  43%|████▎     | 107/250 [03:58<05:16,  2.21s/it]\n",
      "Loss: 0.01350:  43%|████▎     | 107/250 [03:58<05:16,  2.21s/it]\n",
      "Loss: 0.01397:  43%|████▎     | 107/250 [03:58<05:16,  2.21s/it]\n",
      "Loss: 0.01187:  43%|████▎     | 107/250 [03:58<05:16,  2.21s/it]\n",
      "Loss: 0.01108:  43%|████▎     | 107/250 [03:58<05:16,  2.21s/it]\n",
      "Loss: 0.01290:  43%|████▎     | 107/250 [03:58<05:16,  2.21s/it]\n",
      "Loss: 0.01240:  43%|████▎     | 107/250 [03:58<05:16,  2.21s/it]\n",
      "Loss: 0.01536:  43%|████▎     | 107/250 [03:58<05:16,  2.21s/it]\n",
      "Loss: 0.01241:  43%|████▎     | 107/250 [03:58<05:16,  2.21s/it]\n",
      "Loss: 0.01253:  43%|████▎     | 107/250 [03:58<05:16,  2.21s/it]\n",
      "Loss: 0.01253:  43%|████▎     | 108/250 [03:58<05:14,  2.21s/it]\n",
      "Loss: 0.01202:  43%|████▎     | 108/250 [03:59<05:14,  2.21s/it]\n",
      "Loss: 0.01263:  43%|████▎     | 108/250 [03:59<05:14,  2.21s/it]\n",
      "Loss: 0.01444:  43%|████▎     | 108/250 [03:59<05:14,  2.21s/it]\n",
      "Loss: 0.01137:  43%|████▎     | 108/250 [03:59<05:14,  2.21s/it]\n",
      "Loss: 0.01041:  43%|████▎     | 108/250 [03:59<05:14,  2.21s/it]\n",
      "Loss: 0.01293:  43%|████▎     | 108/250 [03:59<05:14,  2.21s/it]\n",
      "Loss: 0.01355:  43%|████▎     | 108/250 [03:59<05:14,  2.21s/it]\n",
      "Loss: 0.01182:  43%|████▎     | 108/250 [03:59<05:14,  2.21s/it]\n",
      "Loss: 0.01463:  43%|████▎     | 108/250 [03:59<05:14,  2.21s/it]\n",
      "Loss: 0.01304:  43%|████▎     | 108/250 [04:00<05:14,  2.21s/it]\n",
      "Loss: 0.01178:  43%|████▎     | 108/250 [04:00<05:14,  2.21s/it]\n",
      "Loss: 0.01168:  43%|████▎     | 108/250 [04:00<05:14,  2.21s/it]\n",
      "Loss: 0.01223:  43%|████▎     | 108/250 [04:00<05:14,  2.21s/it]\n",
      "Loss: 0.01104:  43%|████▎     | 108/250 [04:00<05:14,  2.21s/it]\n",
      "Loss: 0.01178:  43%|████▎     | 108/250 [04:00<05:14,  2.21s/it]\n",
      "Loss: 0.01400:  43%|████▎     | 108/250 [04:00<05:14,  2.21s/it]\n",
      "Loss: 0.01146:  43%|████▎     | 108/250 [04:00<05:14,  2.21s/it]\n",
      "Loss: 0.01371:  43%|████▎     | 108/250 [04:00<05:14,  2.21s/it]\n",
      "Loss: 0.01331:  43%|████▎     | 108/250 [04:00<05:14,  2.21s/it]\n",
      "Loss: 0.01233:  43%|████▎     | 108/250 [04:01<05:14,  2.21s/it]\n",
      "Loss: 0.01280:  43%|████▎     | 108/250 [04:01<05:14,  2.21s/it]\n",
      "Loss: 0.01280:  44%|████▎     | 109/250 [04:01<05:11,  2.21s/it]\n",
      "Loss: 0.01200:  44%|████▎     | 109/250 [04:01<05:11,  2.21s/it]\n",
      "Loss: 0.01268:  44%|████▎     | 109/250 [04:01<05:11,  2.21s/it]\n",
      "Loss: 0.01433:  44%|████▎     | 109/250 [04:01<05:11,  2.21s/it]\n",
      "Loss: 0.01091:  44%|████▎     | 109/250 [04:01<05:11,  2.21s/it]\n",
      "Loss: 0.01063:  44%|████▎     | 109/250 [04:01<05:11,  2.21s/it]\n",
      "Loss: 0.01295:  44%|████▎     | 109/250 [04:01<05:11,  2.21s/it]\n",
      "Loss: 0.01059:  44%|████▎     | 109/250 [04:01<05:11,  2.21s/it]\n",
      "Loss: 0.01606:  44%|████▎     | 109/250 [04:01<05:11,  2.21s/it]\n",
      "Loss: 0.01070:  44%|████▎     | 109/250 [04:02<05:11,  2.21s/it]\n",
      "Loss: 0.01404:  44%|████▎     | 109/250 [04:02<05:11,  2.21s/it]\n",
      "Loss: 0.01405:  44%|████▎     | 109/250 [04:02<05:11,  2.21s/it]\n",
      "Loss: 0.01229:  44%|████▎     | 109/250 [04:02<05:11,  2.21s/it]\n",
      "Loss: 0.01346:  44%|████▎     | 109/250 [04:02<05:11,  2.21s/it]\n",
      "Loss: 0.01458:  44%|████▎     | 109/250 [04:02<05:11,  2.21s/it]\n",
      "Loss: 0.01165:  44%|████▎     | 109/250 [04:02<05:11,  2.21s/it]\n",
      "Loss: 0.01199:  44%|████▎     | 109/250 [04:02<05:11,  2.21s/it]\n",
      "Loss: 0.01339:  44%|████▎     | 109/250 [04:02<05:11,  2.21s/it]\n",
      "Loss: 0.01185:  44%|████▎     | 109/250 [04:03<05:11,  2.21s/it]\n",
      "Loss: 0.01277:  44%|████▎     | 109/250 [04:03<05:11,  2.21s/it]\n",
      "Loss: 0.01210:  44%|████▎     | 109/250 [04:03<05:11,  2.21s/it]\n",
      "Loss: 0.01238:  44%|████▎     | 109/250 [04:03<05:11,  2.21s/it]\n",
      "Loss: 0.01238:  44%|████▍     | 110/250 [04:03<05:08,  2.21s/it]\n",
      "Loss: 0.01183:  44%|████▍     | 110/250 [04:03<05:08,  2.21s/it]\n",
      "Loss: 0.01239:  44%|████▍     | 110/250 [04:03<05:08,  2.21s/it]\n",
      "Loss: 0.01391:  44%|████▍     | 110/250 [04:03<05:08,  2.21s/it]\n",
      "Loss: 0.01106:  44%|████▍     | 110/250 [04:03<05:08,  2.21s/it]\n",
      "Loss: 0.01078:  44%|████▍     | 110/250 [04:03<05:08,  2.21s/it]\n",
      "Loss: 0.01284:  44%|████▍     | 110/250 [04:03<05:08,  2.21s/it]\n",
      "Loss: 0.01078:  44%|████▍     | 110/250 [04:04<05:08,  2.21s/it]\n",
      "Loss: 0.01550:  44%|████▍     | 110/250 [04:04<05:08,  2.21s/it]\n",
      "Loss: 0.01066:  44%|████▍     | 110/250 [04:04<05:08,  2.21s/it]\n",
      "Loss: 0.01409:  44%|████▍     | 110/250 [04:04<05:08,  2.21s/it]\n",
      "Loss: 0.01443:  44%|████▍     | 110/250 [04:04<05:08,  2.21s/it]\n",
      "Loss: 0.01239:  44%|████▍     | 110/250 [04:04<05:08,  2.21s/it]\n",
      "Loss: 0.01341:  44%|████▍     | 110/250 [04:04<05:08,  2.21s/it]\n",
      "Loss: 0.01410:  44%|████▍     | 110/250 [04:04<05:08,  2.21s/it]\n",
      "Loss: 0.01270:  44%|████▍     | 110/250 [04:04<05:08,  2.21s/it]\n",
      "Loss: 0.01136:  44%|████▍     | 110/250 [04:05<05:08,  2.21s/it]\n",
      "Loss: 0.01222:  44%|████▍     | 110/250 [04:05<05:08,  2.21s/it]\n",
      "Loss: 0.01240:  44%|████▍     | 110/250 [04:05<05:08,  2.21s/it]\n",
      "Loss: 0.01532:  44%|████▍     | 110/250 [04:05<05:08,  2.21s/it]\n",
      "Loss: 0.01205:  44%|████▍     | 110/250 [04:05<05:08,  2.21s/it]\n",
      "Loss: 0.01249:  44%|████▍     | 110/250 [04:05<05:08,  2.21s/it]\n",
      "Loss: 0.01249:  44%|████▍     | 111/250 [04:05<05:07,  2.21s/it]\n",
      "Loss: 0.01236:  44%|████▍     | 111/250 [04:05<05:07,  2.21s/it]\n",
      "Loss: 0.01279:  44%|████▍     | 111/250 [04:05<05:07,  2.21s/it]\n",
      "Loss: 0.01388:  44%|████▍     | 111/250 [04:05<05:07,  2.21s/it]\n",
      "Loss: 0.01386:  44%|████▍     | 111/250 [04:05<05:07,  2.21s/it]\n",
      "Loss: 0.01334:  44%|████▍     | 111/250 [04:06<05:07,  2.21s/it]\n",
      "Loss: 0.01243:  44%|████▍     | 111/250 [04:06<05:07,  2.21s/it]\n",
      "Loss: 0.01350:  44%|████▍     | 111/250 [04:06<05:07,  2.21s/it]\n",
      "Loss: 0.01230:  44%|████▍     | 111/250 [04:06<05:07,  2.21s/it]\n",
      "Loss: 0.01454:  44%|████▍     | 111/250 [04:06<05:07,  2.21s/it]\n",
      "Loss: 0.01295:  44%|████▍     | 111/250 [04:06<05:07,  2.21s/it]\n",
      "Loss: 0.01195:  44%|████▍     | 111/250 [04:06<05:07,  2.21s/it]\n",
      "Loss: 0.01337:  44%|████▍     | 111/250 [04:06<05:07,  2.21s/it]\n",
      "Loss: 0.01154:  44%|████▍     | 111/250 [04:06<05:07,  2.21s/it]\n",
      "Loss: 0.01082:  44%|████▍     | 111/250 [04:07<05:07,  2.21s/it]\n",
      "Loss: 0.01146:  44%|████▍     | 111/250 [04:07<05:07,  2.21s/it]\n",
      "Loss: 0.01211:  44%|████▍     | 111/250 [04:07<05:07,  2.21s/it]\n",
      "Loss: 0.01284:  44%|████▍     | 111/250 [04:07<05:07,  2.21s/it]\n",
      "Loss: 0.01247:  44%|████▍     | 111/250 [04:07<05:07,  2.21s/it]\n",
      "Loss: 0.01232:  44%|████▍     | 111/250 [04:07<05:07,  2.21s/it]\n",
      "Loss: 0.01242:  44%|████▍     | 111/250 [04:07<05:07,  2.21s/it]\n",
      "Loss: 0.01242:  45%|████▍     | 112/250 [04:07<05:02,  2.19s/it]\n",
      "Loss: 0.01211:  45%|████▍     | 112/250 [04:07<05:02,  2.19s/it]\n",
      "Loss: 0.01220:  45%|████▍     | 112/250 [04:07<05:02,  2.19s/it]\n",
      "Loss: 0.01223:  45%|████▍     | 112/250 [04:08<05:02,  2.19s/it]\n",
      "Loss: 0.01384:  45%|████▍     | 112/250 [04:08<05:02,  2.19s/it]\n",
      "Loss: 0.00993:  45%|████▍     | 112/250 [04:08<05:02,  2.19s/it]\n",
      "Loss: 0.01089:  45%|████▍     | 112/250 [04:08<05:02,  2.19s/it]\n",
      "Loss: 0.01085:  45%|████▍     | 112/250 [04:08<05:02,  2.19s/it]\n",
      "Loss: 0.01265:  45%|████▍     | 112/250 [04:08<05:02,  2.19s/it]\n",
      "Loss: 0.01135:  45%|████▍     | 112/250 [04:08<05:02,  2.19s/it]\n",
      "Loss: 0.01375:  45%|████▍     | 112/250 [04:08<05:02,  2.19s/it]\n",
      "Loss: 0.01236:  45%|████▍     | 112/250 [04:08<05:02,  2.19s/it]\n",
      "Loss: 0.01358:  45%|████▍     | 112/250 [04:09<05:02,  2.19s/it]\n",
      "Loss: 0.01331:  45%|████▍     | 112/250 [04:09<05:02,  2.19s/it]\n",
      "Loss: 0.01304:  45%|████▍     | 112/250 [04:09<05:02,  2.19s/it]\n",
      "Loss: 0.01322:  45%|████▍     | 112/250 [04:09<05:02,  2.19s/it]\n",
      "Loss: 0.01208:  45%|████▍     | 112/250 [04:09<05:02,  2.19s/it]\n",
      "Loss: 0.01326:  45%|████▍     | 112/250 [04:09<05:02,  2.19s/it]\n",
      "Loss: 0.01216:  45%|████▍     | 112/250 [04:09<05:02,  2.19s/it]\n",
      "Loss: 0.01228:  45%|████▍     | 112/250 [04:09<05:02,  2.19s/it]\n",
      "Loss: 0.01222:  45%|████▍     | 112/250 [04:09<05:02,  2.19s/it]\n",
      "Loss: 0.01299:  45%|████▍     | 112/250 [04:09<05:02,  2.19s/it]\n",
      "Loss: 0.01299:  45%|████▌     | 113/250 [04:09<05:01,  2.20s/it]\n",
      "Loss: 0.01168:  45%|████▌     | 113/250 [04:10<05:01,  2.20s/it]\n",
      "Loss: 0.01261:  45%|████▌     | 113/250 [04:10<05:01,  2.20s/it]\n",
      "Loss: 0.01426:  45%|████▌     | 113/250 [04:10<05:01,  2.20s/it]\n",
      "Loss: 0.01076:  45%|████▌     | 113/250 [04:10<05:01,  2.20s/it]\n",
      "Loss: 0.01040:  45%|████▌     | 113/250 [04:10<05:01,  2.20s/it]\n",
      "Loss: 0.01249:  45%|████▌     | 113/250 [04:10<05:01,  2.20s/it]\n",
      "Loss: 0.01356:  45%|████▌     | 113/250 [04:10<05:01,  2.20s/it]\n",
      "Loss: 0.01267:  45%|████▌     | 113/250 [04:10<05:01,  2.20s/it]\n",
      "Loss: 0.01127:  45%|████▌     | 113/250 [04:10<05:01,  2.20s/it]\n",
      "Loss: 0.01408:  45%|████▌     | 113/250 [04:11<05:01,  2.20s/it]\n",
      "Loss: 0.01255:  45%|████▌     | 113/250 [04:11<05:01,  2.20s/it]\n",
      "Loss: 0.01339:  45%|████▌     | 113/250 [04:11<05:01,  2.20s/it]\n",
      "Loss: 0.01310:  45%|████▌     | 113/250 [04:11<05:01,  2.20s/it]\n",
      "Loss: 0.01110:  45%|████▌     | 113/250 [04:11<05:01,  2.20s/it]\n",
      "Loss: 0.01293:  45%|████▌     | 113/250 [04:11<05:01,  2.20s/it]\n",
      "Loss: 0.01084:  45%|████▌     | 113/250 [04:11<05:01,  2.20s/it]\n",
      "Loss: 0.01308:  45%|████▌     | 113/250 [04:11<05:01,  2.20s/it]\n",
      "Loss: 0.01194:  45%|████▌     | 113/250 [04:11<05:01,  2.20s/it]\n",
      "Loss: 0.01138:  45%|████▌     | 113/250 [04:11<05:01,  2.20s/it]\n",
      "Loss: 0.01219:  45%|████▌     | 113/250 [04:12<05:01,  2.20s/it]\n",
      "Loss: 0.01267:  45%|████▌     | 113/250 [04:12<05:01,  2.20s/it]\n",
      "Loss: 0.01267:  46%|████▌     | 114/250 [04:12<04:59,  2.20s/it]\n",
      "Loss: 0.01165:  46%|████▌     | 114/250 [04:12<04:59,  2.20s/it]\n",
      "Loss: 0.01265:  46%|████▌     | 114/250 [04:12<04:59,  2.20s/it]\n",
      "Loss: 0.01385:  46%|████▌     | 114/250 [04:12<04:59,  2.20s/it]\n",
      "Loss: 0.01052:  46%|████▌     | 114/250 [04:12<04:59,  2.20s/it]\n",
      "Loss: 0.01059:  46%|████▌     | 114/250 [04:12<04:59,  2.20s/it]\n",
      "Loss: 0.01258:  46%|████▌     | 114/250 [04:12<04:59,  2.20s/it]\n",
      "Loss: 0.01033:  46%|████▌     | 114/250 [04:12<04:59,  2.20s/it]\n",
      "Loss: 0.01546:  46%|████▌     | 114/250 [04:12<04:59,  2.20s/it]\n",
      "Loss: 0.01086:  46%|████▌     | 114/250 [04:13<04:59,  2.20s/it]\n",
      "Loss: 0.01395:  46%|████▌     | 114/250 [04:13<04:59,  2.20s/it]\n",
      "Loss: 0.01448:  46%|████▌     | 114/250 [04:13<04:59,  2.20s/it]\n",
      "Loss: 0.01202:  46%|████▌     | 114/250 [04:13<04:59,  2.20s/it]\n",
      "Loss: 0.01297:  46%|████▌     | 114/250 [04:13<04:59,  2.20s/it]\n",
      "Loss: 0.01383:  46%|████▌     | 114/250 [04:13<04:59,  2.20s/it]\n",
      "Loss: 0.01173:  46%|████▌     | 114/250 [04:13<04:59,  2.20s/it]\n",
      "Loss: 0.01108:  46%|████▌     | 114/250 [04:13<04:59,  2.20s/it]\n",
      "Loss: 0.01243:  46%|████▌     | 114/250 [04:13<04:59,  2.20s/it]\n",
      "Loss: 0.01213:  46%|████▌     | 114/250 [04:14<04:59,  2.20s/it]\n",
      "Loss: 0.01528:  46%|████▌     | 114/250 [04:14<04:59,  2.20s/it]\n",
      "Loss: 0.01206:  46%|████▌     | 114/250 [04:14<04:59,  2.20s/it]\n",
      "Loss: 0.01240:  46%|████▌     | 114/250 [04:14<04:59,  2.20s/it]\n",
      "Loss: 0.01240:  46%|████▌     | 115/250 [04:14<04:58,  2.21s/it]\n",
      "Loss: 0.01250:  46%|████▌     | 115/250 [04:14<04:58,  2.21s/it]\n",
      "Loss: 0.01244:  46%|████▌     | 115/250 [04:14<04:58,  2.21s/it]\n",
      "Loss: 0.01179:  46%|████▌     | 115/250 [04:14<04:58,  2.21s/it]\n",
      "Loss: 0.01034:  46%|████▌     | 115/250 [04:14<04:58,  2.21s/it]\n",
      "Loss: 0.01158:  46%|████▌     | 115/250 [04:14<04:58,  2.21s/it]\n",
      "Loss: 0.01220:  46%|████▌     | 115/250 [04:15<04:58,  2.21s/it]\n",
      "Loss: 0.01242:  46%|████▌     | 115/250 [04:15<04:58,  2.21s/it]\n",
      "Loss: 0.01229:  46%|████▌     | 115/250 [04:15<04:58,  2.21s/it]\n",
      "Loss: 0.01292:  46%|████▌     | 115/250 [04:15<04:58,  2.21s/it]\n",
      "Loss: 0.01120:  46%|████▌     | 115/250 [04:15<04:58,  2.21s/it]\n",
      "Loss: 0.01434:  46%|████▌     | 115/250 [04:15<04:58,  2.21s/it]\n",
      "Loss: 0.01225:  46%|████▌     | 115/250 [04:15<04:58,  2.21s/it]\n",
      "Loss: 0.01350:  46%|████▌     | 115/250 [04:15<04:58,  2.21s/it]\n",
      "Loss: 0.01337:  46%|████▌     | 115/250 [04:15<04:58,  2.21s/it]\n",
      "Loss: 0.01297:  46%|████▌     | 115/250 [04:16<04:58,  2.21s/it]\n",
      "Loss: 0.01314:  46%|████▌     | 115/250 [04:16<04:58,  2.21s/it]\n",
      "Loss: 0.01166:  46%|████▌     | 115/250 [04:16<04:58,  2.21s/it]\n",
      "Loss: 0.01396:  46%|████▌     | 115/250 [04:16<04:58,  2.21s/it]\n",
      "Loss: 0.01165:  46%|████▌     | 115/250 [04:16<04:58,  2.21s/it]\n",
      "Loss: 0.01280:  46%|████▌     | 115/250 [04:16<04:58,  2.21s/it]\n",
      "Loss: 0.01198:  46%|████▌     | 115/250 [04:16<04:58,  2.21s/it]\n",
      "Loss: 0.01234:  46%|████▌     | 115/250 [04:16<04:58,  2.21s/it]\n",
      "Loss: 0.01234:  46%|████▋     | 116/250 [04:16<05:00,  2.24s/it]\n",
      "Loss: 0.01156:  46%|████▋     | 116/250 [04:16<05:00,  2.24s/it]\n",
      "Loss: 0.01285:  46%|████▋     | 116/250 [04:16<05:00,  2.24s/it]\n",
      "Loss: 0.01207:  46%|████▋     | 116/250 [04:17<05:00,  2.24s/it]\n",
      "Loss: 0.01369:  46%|████▋     | 116/250 [04:17<05:00,  2.24s/it]\n",
      "Loss: 0.00995:  46%|████▋     | 116/250 [04:17<05:00,  2.24s/it]\n",
      "Loss: 0.01281:  46%|████▋     | 116/250 [04:17<05:00,  2.24s/it]\n",
      "Loss: 0.01238:  46%|████▋     | 116/250 [04:17<05:00,  2.24s/it]\n",
      "Loss: 0.01301:  46%|████▋     | 116/250 [04:17<05:00,  2.24s/it]\n",
      "Loss: 0.01115:  46%|████▋     | 116/250 [04:17<05:00,  2.24s/it]\n",
      "Loss: 0.01384:  46%|████▋     | 116/250 [04:17<05:00,  2.24s/it]\n",
      "Loss: 0.01216:  46%|████▋     | 116/250 [04:17<05:00,  2.24s/it]\n",
      "Loss: 0.01234:  46%|████▋     | 116/250 [04:18<05:00,  2.24s/it]\n",
      "Loss: 0.01199:  46%|████▋     | 116/250 [04:18<05:00,  2.24s/it]\n",
      "Loss: 0.01042:  46%|████▋     | 116/250 [04:18<05:00,  2.24s/it]\n",
      "Loss: 0.01182:  46%|████▋     | 116/250 [04:18<05:00,  2.24s/it]\n",
      "Loss: 0.01167:  46%|████▋     | 116/250 [04:18<05:00,  2.24s/it]\n",
      "Loss: 0.01271:  46%|████▋     | 116/250 [04:18<05:00,  2.24s/it]\n",
      "Loss: 0.01286:  46%|████▋     | 116/250 [04:18<05:00,  2.24s/it]\n",
      "Loss: 0.01134:  46%|████▋     | 116/250 [04:18<05:00,  2.24s/it]\n",
      "Loss: 0.01228:  46%|████▋     | 116/250 [04:18<05:00,  2.24s/it]\n",
      "Loss: 0.01228:  47%|████▋     | 117/250 [04:18<04:55,  2.22s/it]\n",
      "Loss: 0.01141:  47%|████▋     | 117/250 [04:18<04:55,  2.22s/it]\n",
      "Loss: 0.01247:  47%|████▋     | 117/250 [04:19<04:55,  2.22s/it]\n",
      "Loss: 0.01211:  47%|████▋     | 117/250 [04:19<04:55,  2.22s/it]\n",
      "Loss: 0.01395:  47%|████▋     | 117/250 [04:19<04:55,  2.22s/it]\n",
      "Loss: 0.01325:  47%|████▋     | 117/250 [04:19<04:55,  2.22s/it]\n",
      "Loss: 0.01247:  47%|████▋     | 117/250 [04:19<04:55,  2.22s/it]\n",
      "Loss: 0.01344:  47%|████▋     | 117/250 [04:19<04:55,  2.22s/it]\n",
      "Loss: 0.01270:  47%|████▋     | 117/250 [04:19<04:55,  2.22s/it]\n",
      "Loss: 0.01146:  47%|████▋     | 117/250 [04:19<04:55,  2.22s/it]\n",
      "Loss: 0.01384:  47%|████▋     | 117/250 [04:19<04:55,  2.22s/it]\n",
      "Loss: 0.01242:  47%|████▋     | 117/250 [04:20<04:55,  2.22s/it]\n",
      "Loss: 0.01353:  47%|████▋     | 117/250 [04:20<04:55,  2.22s/it]\n",
      "Loss: 0.01310:  47%|████▋     | 117/250 [04:20<04:55,  2.22s/it]\n",
      "Loss: 0.01271:  47%|████▋     | 117/250 [04:20<04:55,  2.22s/it]\n",
      "Loss: 0.01282:  47%|████▋     | 117/250 [04:20<04:55,  2.22s/it]\n",
      "Loss: 0.01251:  47%|████▋     | 117/250 [04:20<04:55,  2.22s/it]\n",
      "Loss: 0.01271:  47%|████▋     | 117/250 [04:20<04:55,  2.22s/it]\n",
      "Loss: 0.01199:  47%|████▋     | 117/250 [04:20<04:55,  2.22s/it]\n",
      "Loss: 0.01108:  47%|████▋     | 117/250 [04:20<04:55,  2.22s/it]\n",
      "Loss: 0.01195:  47%|████▋     | 117/250 [04:21<04:55,  2.22s/it]\n",
      "Loss: 0.01264:  47%|████▋     | 117/250 [04:21<04:55,  2.22s/it]\n",
      "Loss: 0.01264:  47%|████▋     | 118/250 [04:21<04:51,  2.21s/it]\n",
      "Loss: 0.01157:  47%|████▋     | 118/250 [04:21<04:51,  2.21s/it]\n",
      "Loss: 0.01256:  47%|████▋     | 118/250 [04:21<04:51,  2.21s/it]\n",
      "Loss: 0.01407:  47%|████▋     | 118/250 [04:21<04:51,  2.21s/it]\n",
      "Loss: 0.01073:  47%|████▋     | 118/250 [04:21<04:51,  2.21s/it]\n",
      "Loss: 0.01084:  47%|████▋     | 118/250 [04:21<04:51,  2.21s/it]\n",
      "Loss: 0.01288:  47%|████▋     | 118/250 [04:21<04:51,  2.21s/it]\n",
      "Loss: 0.01061:  47%|████▋     | 118/250 [04:21<04:51,  2.21s/it]\n",
      "Loss: 0.01556:  47%|████▋     | 118/250 [04:21<04:51,  2.21s/it]\n",
      "Loss: 0.01401:  47%|████▋     | 118/250 [04:22<04:51,  2.21s/it]\n",
      "Loss: 0.01258:  47%|████▋     | 118/250 [04:22<04:51,  2.21s/it]\n",
      "Loss: 0.01163:  47%|████▋     | 118/250 [04:22<04:51,  2.21s/it]\n",
      "Loss: 0.01150:  47%|████▋     | 118/250 [04:22<04:51,  2.21s/it]\n",
      "Loss: 0.01193:  47%|████▋     | 118/250 [04:22<04:51,  2.21s/it]\n",
      "Loss: 0.01067:  47%|████▋     | 118/250 [04:22<04:51,  2.21s/it]\n",
      "Loss: 0.01118:  47%|████▋     | 118/250 [04:22<04:51,  2.21s/it]\n",
      "Loss: 0.01192:  47%|████▋     | 118/250 [04:22<04:51,  2.21s/it]\n",
      "Loss: 0.01290:  47%|████▋     | 118/250 [04:22<04:51,  2.21s/it]\n",
      "Loss: 0.01157:  47%|████▋     | 118/250 [04:23<04:51,  2.21s/it]\n",
      "Loss: 0.01270:  47%|████▋     | 118/250 [04:23<04:51,  2.21s/it]\n",
      "Loss: 0.01229:  47%|████▋     | 118/250 [04:23<04:51,  2.21s/it]\n",
      "Loss: 0.01264:  47%|████▋     | 118/250 [04:23<04:51,  2.21s/it]\n",
      "Loss: 0.01264:  48%|████▊     | 119/250 [04:23<04:51,  2.23s/it]\n",
      "Loss: 0.01183:  48%|████▊     | 119/250 [04:23<04:51,  2.23s/it]\n",
      "Loss: 0.01223:  48%|████▊     | 119/250 [04:23<04:51,  2.23s/it]\n",
      "Loss: 0.01414:  48%|████▊     | 119/250 [04:23<04:51,  2.23s/it]\n",
      "Loss: 0.01062:  48%|████▊     | 119/250 [04:23<04:51,  2.23s/it]\n",
      "Loss: 0.01054:  48%|████▊     | 119/250 [04:23<04:51,  2.23s/it]\n",
      "Loss: 0.01256:  48%|████▊     | 119/250 [04:23<04:51,  2.23s/it]\n",
      "Loss: 0.01009:  48%|████▊     | 119/250 [04:24<04:51,  2.23s/it]\n",
      "Loss: 0.01574:  48%|████▊     | 119/250 [04:24<04:51,  2.23s/it]\n",
      "Loss: 0.01068:  48%|████▊     | 119/250 [04:24<04:51,  2.23s/it]\n",
      "Loss: 0.01346:  48%|████▊     | 119/250 [04:24<04:51,  2.23s/it]\n",
      "Loss: 0.01425:  48%|████▊     | 119/250 [04:24<04:51,  2.23s/it]\n",
      "Loss: 0.01183:  48%|████▊     | 119/250 [04:24<04:51,  2.23s/it]\n",
      "Loss: 0.01358:  48%|████▊     | 119/250 [04:24<04:51,  2.23s/it]\n",
      "Loss: 0.01378:  48%|████▊     | 119/250 [04:24<04:51,  2.23s/it]\n",
      "Loss: 0.01189:  48%|████▊     | 119/250 [04:24<04:51,  2.23s/it]\n",
      "Loss: 0.01092:  48%|████▊     | 119/250 [04:25<04:51,  2.23s/it]\n",
      "Loss: 0.01238:  48%|████▊     | 119/250 [04:25<04:51,  2.23s/it]\n",
      "Loss: 0.01215:  48%|████▊     | 119/250 [04:25<04:51,  2.23s/it]\n",
      "Loss: 0.01499:  48%|████▊     | 119/250 [04:25<04:51,  2.23s/it]\n",
      "Loss: 0.01174:  48%|████▊     | 119/250 [04:25<04:51,  2.23s/it]\n",
      "Loss: 0.01247:  48%|████▊     | 119/250 [04:25<04:51,  2.23s/it]\n",
      "Loss: 0.01247:  48%|████▊     | 120/250 [04:25<04:48,  2.22s/it]\n",
      "Loss: 0.01175:  48%|████▊     | 120/250 [04:25<04:48,  2.22s/it]\n",
      "Loss: 0.01225:  48%|████▊     | 120/250 [04:25<04:48,  2.22s/it]\n",
      "Loss: 0.01394:  48%|████▊     | 120/250 [04:25<04:48,  2.22s/it]\n",
      "Loss: 0.01049:  48%|████▊     | 120/250 [04:25<04:48,  2.22s/it]\n",
      "Loss: 0.01031:  48%|████▊     | 120/250 [04:26<04:48,  2.22s/it]\n",
      "Loss: 0.01266:  48%|████▊     | 120/250 [04:26<04:48,  2.22s/it]\n",
      "Loss: 0.01314:  48%|████▊     | 120/250 [04:26<04:48,  2.22s/it]\n",
      "Loss: 0.01174:  48%|████▊     | 120/250 [04:26<04:48,  2.22s/it]\n",
      "Loss: 0.01397:  48%|████▊     | 120/250 [04:26<04:48,  2.22s/it]\n",
      "Loss: 0.01252:  48%|████▊     | 120/250 [04:26<04:48,  2.22s/it]\n",
      "Loss: 0.01176:  48%|████▊     | 120/250 [04:26<04:48,  2.22s/it]\n",
      "Loss: 0.01151:  48%|████▊     | 120/250 [04:26<04:48,  2.22s/it]\n",
      "Loss: 0.01218:  48%|████▊     | 120/250 [04:26<04:48,  2.22s/it]\n",
      "Loss: 0.01076:  48%|████▊     | 120/250 [04:27<04:48,  2.22s/it]\n",
      "Loss: 0.01174:  48%|████▊     | 120/250 [04:27<04:48,  2.22s/it]\n",
      "Loss: 0.01199:  48%|████▊     | 120/250 [04:27<04:48,  2.22s/it]\n",
      "Loss: 0.01362:  48%|████▊     | 120/250 [04:27<04:48,  2.22s/it]\n",
      "Loss: 0.01148:  48%|████▊     | 120/250 [04:27<04:48,  2.22s/it]\n",
      "Loss: 0.01257:  48%|████▊     | 120/250 [04:27<04:48,  2.22s/it]\n",
      "Loss: 0.01210:  48%|████▊     | 120/250 [04:27<04:48,  2.22s/it]\n",
      "Loss: 0.01245:  48%|████▊     | 120/250 [04:27<04:48,  2.22s/it]\n",
      "Loss: 0.01245:  48%|████▊     | 121/250 [04:27<04:45,  2.22s/it]\n",
      "Loss: 0.01233:  48%|████▊     | 121/250 [04:27<04:45,  2.22s/it]\n",
      "Loss: 0.01177:  48%|████▊     | 121/250 [04:27<04:45,  2.22s/it]\n",
      "Loss: 0.01205:  48%|████▊     | 121/250 [04:28<04:45,  2.22s/it]\n",
      "Loss: 0.01362:  48%|████▊     | 121/250 [04:28<04:45,  2.22s/it]\n",
      "Loss: 0.01424:  48%|████▊     | 121/250 [04:28<04:45,  2.22s/it]\n",
      "Loss: 0.01072:  48%|████▊     | 121/250 [04:28<04:45,  2.22s/it]\n",
      "Loss: 0.01084:  48%|████▊     | 121/250 [04:28<04:45,  2.22s/it]\n",
      "Loss: 0.01236:  48%|████▊     | 121/250 [04:28<04:45,  2.22s/it]\n",
      "Loss: 0.01121:  48%|████▊     | 121/250 [04:28<04:45,  2.22s/it]\n",
      "Loss: 0.01154:  48%|████▊     | 121/250 [04:28<04:45,  2.22s/it]\n",
      "Loss: 0.01216:  48%|████▊     | 121/250 [04:28<04:45,  2.22s/it]\n",
      "Loss: 0.01209:  48%|████▊     | 121/250 [04:29<04:45,  2.22s/it]\n",
      "Loss: 0.01169:  48%|████▊     | 121/250 [04:29<04:45,  2.22s/it]\n",
      "Loss: 0.01003:  48%|████▊     | 121/250 [04:29<04:45,  2.22s/it]\n",
      "Loss: 0.01102:  48%|████▊     | 121/250 [04:29<04:45,  2.22s/it]\n",
      "Loss: 0.01180:  48%|████▊     | 121/250 [04:29<04:45,  2.22s/it]\n",
      "Loss: 0.01267:  48%|████▊     | 121/250 [04:29<04:45,  2.22s/it]\n",
      "Loss: 0.01265:  48%|████▊     | 121/250 [04:29<04:45,  2.22s/it]\n",
      "Loss: 0.01095:  48%|████▊     | 121/250 [04:29<04:45,  2.22s/it]\n",
      "Loss: 0.01251:  48%|████▊     | 121/250 [04:29<04:45,  2.22s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 55.83it/s]\u001b[A\n",
      "Loss: 0.01251:  49%|████▉     | 122/250 [04:29<04:42,  2.20s/it]\n",
      "Loss: 0.01153:  49%|████▉     | 122/250 [04:30<04:42,  2.20s/it]\n",
      "Loss: 0.01224:  49%|████▉     | 122/250 [04:30<04:42,  2.20s/it]\n",
      "Loss: 0.01385:  49%|████▉     | 122/250 [04:30<04:42,  2.20s/it]\n",
      "Loss: 0.01039:  49%|████▉     | 122/250 [04:30<04:42,  2.20s/it]\n",
      "Loss: 0.01013:  49%|████▉     | 122/250 [04:30<04:42,  2.20s/it]\n",
      "Loss: 0.01263:  49%|████▉     | 122/250 [04:30<04:42,  2.20s/it]\n",
      "Loss: 0.01002:  49%|████▉     | 122/250 [04:30<04:42,  2.20s/it]\n",
      "Loss: 0.01527:  49%|████▉     | 122/250 [04:30<04:42,  2.20s/it]\n",
      "Loss: 0.01064:  49%|████▉     | 122/250 [04:30<04:42,  2.20s/it]\n",
      "Loss: 0.01356:  49%|████▉     | 122/250 [04:30<04:42,  2.20s/it]\n",
      "Loss: 0.01447:  49%|████▉     | 122/250 [04:31<04:42,  2.20s/it]\n",
      "Loss: 0.01240:  49%|████▉     | 122/250 [04:31<04:42,  2.20s/it]\n",
      "Loss: 0.01295:  49%|████▉     | 122/250 [04:31<04:42,  2.20s/it]\n",
      "Loss: 0.01368:  49%|████▉     | 122/250 [04:31<04:42,  2.20s/it]\n",
      "Loss: 0.01153:  49%|████▉     | 122/250 [04:31<04:42,  2.20s/it]\n",
      "Loss: 0.01062:  49%|████▉     | 122/250 [04:31<04:42,  2.20s/it]\n",
      "Loss: 0.01229:  49%|████▉     | 122/250 [04:31<04:42,  2.20s/it]\n",
      "Loss: 0.01234:  49%|████▉     | 122/250 [04:31<04:42,  2.20s/it]\n",
      "Loss: 0.01483:  49%|████▉     | 122/250 [04:31<04:42,  2.20s/it]\n",
      "Loss: 0.01165:  49%|████▉     | 122/250 [04:32<04:42,  2.20s/it]\n",
      "Loss: 0.01261:  49%|████▉     | 122/250 [04:32<04:42,  2.20s/it]\n",
      "Loss: 0.01261:  49%|████▉     | 123/250 [04:32<04:40,  2.21s/it]\n",
      "Loss: 0.01238:  49%|████▉     | 123/250 [04:32<04:40,  2.21s/it]\n",
      "Loss: 0.01181:  49%|████▉     | 123/250 [04:32<04:40,  2.21s/it]\n",
      "Loss: 0.01194:  49%|████▉     | 123/250 [04:32<04:40,  2.21s/it]\n",
      "Loss: 0.01194:  49%|████▉     | 123/250 [04:32<04:40,  2.21s/it]\n",
      "Loss: 0.01006:  49%|████▉     | 123/250 [04:32<04:40,  2.21s/it]\n",
      "Loss: 0.01076:  49%|████▉     | 123/250 [04:32<04:40,  2.21s/it]\n",
      "Loss: 0.01106:  49%|████▉     | 123/250 [04:32<04:40,  2.21s/it]\n",
      "Loss: 0.01273:  49%|████▉     | 123/250 [04:33<04:40,  2.21s/it]\n",
      "Loss: 0.01089:  49%|████▉     | 123/250 [04:33<04:40,  2.21s/it]\n",
      "Loss: 0.01336:  49%|████▉     | 123/250 [04:33<04:40,  2.21s/it]\n",
      "Loss: 0.01190:  49%|████▉     | 123/250 [04:33<04:40,  2.21s/it]\n",
      "Loss: 0.01223:  49%|████▉     | 123/250 [04:33<04:40,  2.21s/it]\n",
      "Loss: 0.01199:  49%|████▉     | 123/250 [04:33<04:40,  2.21s/it]\n",
      "Loss: 0.00997:  49%|████▉     | 123/250 [04:33<04:40,  2.21s/it]\n",
      "Loss: 0.01094:  49%|████▉     | 123/250 [04:33<04:40,  2.21s/it]\n",
      "Loss: 0.01101:  49%|████▉     | 123/250 [04:33<04:40,  2.21s/it]\n",
      "Loss: 0.01250:  49%|████▉     | 123/250 [04:33<04:40,  2.21s/it]\n",
      "Loss: 0.01192:  49%|████▉     | 123/250 [04:34<04:40,  2.21s/it]\n",
      "Loss: 0.01295:  49%|████▉     | 123/250 [04:34<04:40,  2.21s/it]\n",
      "Loss: 0.01191:  49%|████▉     | 123/250 [04:34<04:40,  2.21s/it]\n",
      "Loss: 0.01225:  49%|████▉     | 123/250 [04:34<04:40,  2.21s/it]\n",
      "Loss: 0.01225:  50%|████▉     | 124/250 [04:34<04:38,  2.21s/it]\n",
      "Loss: 0.01134:  50%|████▉     | 124/250 [04:34<04:38,  2.21s/it]\n",
      "Loss: 0.01217:  50%|████▉     | 124/250 [04:34<04:38,  2.21s/it]\n",
      "Loss: 0.01357:  50%|████▉     | 124/250 [04:34<04:38,  2.21s/it]\n",
      "Loss: 0.01040:  50%|████▉     | 124/250 [04:34<04:38,  2.21s/it]\n",
      "Loss: 0.01032:  50%|████▉     | 124/250 [04:34<04:38,  2.21s/it]\n",
      "Loss: 0.01233:  50%|████▉     | 124/250 [04:34<04:38,  2.21s/it]\n",
      "Loss: 0.01011:  50%|████▉     | 124/250 [04:35<04:38,  2.21s/it]\n",
      "Loss: 0.01552:  50%|████▉     | 124/250 [04:35<04:38,  2.21s/it]\n",
      "Loss: 0.01017:  50%|████▉     | 124/250 [04:35<04:38,  2.21s/it]\n",
      "Loss: 0.01360:  50%|████▉     | 124/250 [04:35<04:38,  2.21s/it]\n",
      "Loss: 0.01392:  50%|████▉     | 124/250 [04:35<04:38,  2.21s/it]\n",
      "Loss: 0.01219:  50%|████▉     | 124/250 [04:35<04:38,  2.21s/it]\n",
      "Loss: 0.01316:  50%|████▉     | 124/250 [04:35<04:38,  2.21s/it]\n",
      "Loss: 0.01053:  50%|████▉     | 124/250 [04:35<04:38,  2.21s/it]\n",
      "Loss: 0.01107:  50%|████▉     | 124/250 [04:35<04:38,  2.21s/it]\n",
      "Loss: 0.01351:  50%|████▉     | 124/250 [04:36<04:38,  2.21s/it]\n",
      "Loss: 0.01099:  50%|████▉     | 124/250 [04:36<04:38,  2.21s/it]\n",
      "Loss: 0.01339:  50%|████▉     | 124/250 [04:36<04:38,  2.21s/it]\n",
      "Loss: 0.01290:  50%|████▉     | 124/250 [04:36<04:38,  2.21s/it]\n",
      "Loss: 0.01178:  50%|████▉     | 124/250 [04:36<04:38,  2.21s/it]\n",
      "Loss: 0.01257:  50%|████▉     | 124/250 [04:36<04:38,  2.21s/it]\n",
      "Loss: 0.01257:  50%|█████     | 125/250 [04:36<04:36,  2.21s/it]\n",
      "Loss: 0.01151:  50%|█████     | 125/250 [04:36<04:36,  2.21s/it]\n",
      "Loss: 0.01210:  50%|█████     | 125/250 [04:36<04:36,  2.21s/it]\n",
      "Loss: 0.01406:  50%|█████     | 125/250 [04:36<04:36,  2.21s/it]\n",
      "Loss: 0.01097:  50%|█████     | 125/250 [04:36<04:36,  2.21s/it]\n",
      "Loss: 0.01001:  50%|█████     | 125/250 [04:37<04:36,  2.21s/it]\n",
      "Loss: 0.01232:  50%|█████     | 125/250 [04:37<04:36,  2.21s/it]\n",
      "Loss: 0.01013:  50%|█████     | 125/250 [04:37<04:36,  2.21s/it]\n",
      "Loss: 0.01543:  50%|█████     | 125/250 [04:37<04:36,  2.21s/it]\n",
      "Loss: 0.01056:  50%|█████     | 125/250 [04:37<04:36,  2.21s/it]\n",
      "Loss: 0.01357:  50%|█████     | 125/250 [04:37<04:36,  2.21s/it]\n",
      "Loss: 0.01382:  50%|█████     | 125/250 [04:37<04:36,  2.21s/it]\n",
      "Loss: 0.01191:  50%|█████     | 125/250 [04:37<04:36,  2.21s/it]\n",
      "Loss: 0.01394:  50%|█████     | 125/250 [04:37<04:36,  2.21s/it]\n",
      "Loss: 0.01046:  50%|█████     | 125/250 [04:38<04:36,  2.21s/it]\n",
      "Loss: 0.01093:  50%|█████     | 125/250 [04:38<04:36,  2.21s/it]\n",
      "Loss: 0.01346:  50%|█████     | 125/250 [04:38<04:36,  2.21s/it]\n",
      "Loss: 0.01128:  50%|█████     | 125/250 [04:38<04:36,  2.21s/it]\n",
      "Loss: 0.01319:  50%|█████     | 125/250 [04:38<04:36,  2.21s/it]\n",
      "Loss: 0.01318:  50%|█████     | 125/250 [04:38<04:36,  2.21s/it]\n",
      "Loss: 0.01202:  50%|█████     | 125/250 [04:38<04:36,  2.21s/it]\n",
      "Loss: 0.01297:  50%|█████     | 125/250 [04:38<04:36,  2.21s/it]\n",
      "Loss: 0.01297:  50%|█████     | 126/250 [04:38<04:33,  2.20s/it]\n",
      "Loss: 0.01120:  50%|█████     | 126/250 [04:38<04:33,  2.20s/it]\n",
      "Loss: 0.01209:  50%|█████     | 126/250 [04:38<04:33,  2.20s/it]\n",
      "Loss: 0.01338:  50%|█████     | 126/250 [04:39<04:33,  2.20s/it]\n",
      "Loss: 0.01071:  50%|█████     | 126/250 [04:39<04:33,  2.20s/it]\n",
      "Loss: 0.01049:  50%|█████     | 126/250 [04:39<04:33,  2.20s/it]\n",
      "Loss: 0.01253:  50%|█████     | 126/250 [04:39<04:33,  2.20s/it]\n",
      "Loss: 0.01018:  50%|█████     | 126/250 [04:39<04:33,  2.20s/it]\n",
      "Loss: 0.01565:  50%|█████     | 126/250 [04:39<04:33,  2.20s/it]\n",
      "Loss: 0.01045:  50%|█████     | 126/250 [04:39<04:33,  2.20s/it]\n",
      "Loss: 0.01335:  50%|█████     | 126/250 [04:39<04:33,  2.20s/it]\n",
      "Loss: 0.01368:  50%|█████     | 126/250 [04:39<04:33,  2.20s/it]\n",
      "Loss: 0.01198:  50%|█████     | 126/250 [04:39<04:33,  2.20s/it]\n",
      "Loss: 0.01298:  50%|█████     | 126/250 [04:40<04:33,  2.20s/it]\n",
      "Loss: 0.01420:  50%|█████     | 126/250 [04:40<04:33,  2.20s/it]\n",
      "Loss: 0.01147:  50%|█████     | 126/250 [04:40<04:33,  2.20s/it]\n",
      "Loss: 0.01086:  50%|█████     | 126/250 [04:40<04:33,  2.20s/it]\n",
      "Loss: 0.01100:  50%|█████     | 126/250 [04:40<04:33,  2.20s/it]\n",
      "Loss: 0.01344:  50%|█████     | 126/250 [04:40<04:33,  2.20s/it]\n",
      "Loss: 0.01321:  50%|█████     | 126/250 [04:40<04:33,  2.20s/it]\n",
      "Loss: 0.01199:  50%|█████     | 126/250 [04:40<04:33,  2.20s/it]\n",
      "Loss: 0.01234:  50%|█████     | 126/250 [04:40<04:33,  2.20s/it]\n",
      "Loss: 0.01234:  51%|█████     | 127/250 [04:40<04:31,  2.21s/it]\n",
      "Loss: 0.01140:  51%|█████     | 127/250 [04:41<04:31,  2.21s/it]\n",
      "Loss: 0.01205:  51%|█████     | 127/250 [04:41<04:31,  2.21s/it]\n",
      "Loss: 0.01354:  51%|█████     | 127/250 [04:41<04:31,  2.21s/it]\n",
      "Loss: 0.01070:  51%|█████     | 127/250 [04:41<04:31,  2.21s/it]\n",
      "Loss: 0.01030:  51%|█████     | 127/250 [04:41<04:31,  2.21s/it]\n",
      "Loss: 0.01236:  51%|█████     | 127/250 [04:41<04:31,  2.21s/it]\n",
      "Loss: 0.01299:  51%|█████     | 127/250 [04:41<04:31,  2.21s/it]\n",
      "Loss: 0.01229:  51%|█████     | 127/250 [04:41<04:31,  2.21s/it]\n",
      "Loss: 0.01419:  51%|█████     | 127/250 [04:41<04:31,  2.21s/it]\n",
      "Loss: 0.01224:  51%|█████     | 127/250 [04:42<04:31,  2.21s/it]\n",
      "Loss: 0.01131:  51%|█████     | 127/250 [04:42<04:31,  2.21s/it]\n",
      "Loss: 0.01174:  51%|█████     | 127/250 [04:42<04:31,  2.21s/it]\n",
      "Loss: 0.01195:  51%|█████     | 127/250 [04:42<04:31,  2.21s/it]\n",
      "Loss: 0.01033:  51%|█████     | 127/250 [04:42<04:31,  2.21s/it]\n",
      "Loss: 0.01100:  51%|█████     | 127/250 [04:42<04:31,  2.21s/it]\n",
      "Loss: 0.01322:  51%|█████     | 127/250 [04:42<04:31,  2.21s/it]\n",
      "Loss: 0.01098:  51%|█████     | 127/250 [04:42<04:31,  2.21s/it]\n",
      "Loss: 0.01326:  51%|█████     | 127/250 [04:42<04:31,  2.21s/it]\n",
      "Loss: 0.01302:  51%|█████     | 127/250 [04:42<04:31,  2.21s/it]\n",
      "Loss: 0.01197:  51%|█████     | 127/250 [04:43<04:31,  2.21s/it]\n",
      "Loss: 0.01268:  51%|█████     | 127/250 [04:43<04:31,  2.21s/it]\n",
      "Loss: 0.01268:  51%|█████     | 128/250 [04:43<04:30,  2.21s/it]\n",
      "Loss: 0.01118:  51%|█████     | 128/250 [04:43<04:30,  2.21s/it]\n",
      "Loss: 0.01181:  51%|█████     | 128/250 [04:43<04:30,  2.21s/it]\n",
      "Loss: 0.01345:  51%|█████     | 128/250 [04:43<04:30,  2.21s/it]\n",
      "Loss: 0.01078:  51%|█████     | 128/250 [04:43<04:30,  2.21s/it]\n",
      "Loss: 0.01319:  51%|█████     | 128/250 [04:43<04:30,  2.21s/it]\n",
      "Loss: 0.01224:  51%|█████     | 128/250 [04:43<04:30,  2.21s/it]\n",
      "Loss: 0.01337:  51%|█████     | 128/250 [04:43<04:30,  2.21s/it]\n",
      "Loss: 0.01144:  51%|█████     | 128/250 [04:44<04:30,  2.21s/it]\n",
      "Loss: 0.01401:  51%|█████     | 128/250 [04:44<04:30,  2.21s/it]\n",
      "Loss: 0.01240:  51%|█████     | 128/250 [04:44<04:30,  2.21s/it]\n",
      "Loss: 0.01128:  51%|█████     | 128/250 [04:44<04:30,  2.21s/it]\n",
      "Loss: 0.01135:  51%|█████     | 128/250 [04:44<04:30,  2.21s/it]\n",
      "Loss: 0.01165:  51%|█████     | 128/250 [04:44<04:30,  2.21s/it]\n",
      "Loss: 0.01057:  51%|█████     | 128/250 [04:44<04:30,  2.21s/it]\n",
      "Loss: 0.01105:  51%|█████     | 128/250 [04:44<04:30,  2.21s/it]\n",
      "Loss: 0.01333:  51%|█████     | 128/250 [04:44<04:30,  2.21s/it]\n",
      "Loss: 0.01076:  51%|█████     | 128/250 [04:44<04:30,  2.21s/it]\n",
      "Loss: 0.01334:  51%|█████     | 128/250 [04:45<04:30,  2.21s/it]\n",
      "Loss: 0.01263:  51%|█████     | 128/250 [04:45<04:30,  2.21s/it]\n",
      "Loss: 0.01191:  51%|█████     | 128/250 [04:45<04:30,  2.21s/it]\n",
      "Loss: 0.01221:  51%|█████     | 128/250 [04:45<04:30,  2.21s/it]\n",
      "Loss: 0.01221:  52%|█████▏    | 129/250 [04:45<04:26,  2.20s/it]\n",
      "Loss: 0.01134:  52%|█████▏    | 129/250 [04:45<04:26,  2.20s/it]\n",
      "Loss: 0.01196:  52%|█████▏    | 129/250 [04:45<04:26,  2.20s/it]\n",
      "Loss: 0.01336:  52%|█████▏    | 129/250 [04:45<04:26,  2.20s/it]\n",
      "Loss: 0.01038:  52%|█████▏    | 129/250 [04:45<04:26,  2.20s/it]\n",
      "Loss: 0.01003:  52%|█████▏    | 129/250 [04:45<04:26,  2.20s/it]\n",
      "Loss: 0.01232:  52%|█████▏    | 129/250 [04:45<04:26,  2.20s/it]\n",
      "Loss: 0.00993:  52%|█████▏    | 129/250 [04:46<04:26,  2.20s/it]\n",
      "Loss: 0.01502:  52%|█████▏    | 129/250 [04:46<04:26,  2.20s/it]\n",
      "Loss: 0.01066:  52%|█████▏    | 129/250 [04:46<04:26,  2.20s/it]\n",
      "Loss: 0.01353:  52%|█████▏    | 129/250 [04:46<04:26,  2.20s/it]\n",
      "Loss: 0.01355:  52%|█████▏    | 129/250 [04:46<04:26,  2.20s/it]\n",
      "Loss: 0.01152:  52%|█████▏    | 129/250 [04:46<04:26,  2.20s/it]\n",
      "Loss: 0.01249:  52%|█████▏    | 129/250 [04:46<04:26,  2.20s/it]\n",
      "Loss: 0.01387:  52%|█████▏    | 129/250 [04:46<04:26,  2.20s/it]\n",
      "Loss: 0.01130:  52%|█████▏    | 129/250 [04:46<04:26,  2.20s/it]\n",
      "Loss: 0.01061:  52%|█████▏    | 129/250 [04:47<04:26,  2.20s/it]\n",
      "Loss: 0.01199:  52%|█████▏    | 129/250 [04:47<04:26,  2.20s/it]\n",
      "Loss: 0.01226:  52%|█████▏    | 129/250 [04:47<04:26,  2.20s/it]\n",
      "Loss: 0.01456:  52%|█████▏    | 129/250 [04:47<04:26,  2.20s/it]\n",
      "Loss: 0.01208:  52%|█████▏    | 129/250 [04:47<04:26,  2.20s/it]\n",
      "Loss: 0.01235:  52%|█████▏    | 129/250 [04:47<04:26,  2.20s/it]\n",
      "Loss: 0.01235:  52%|█████▏    | 130/250 [04:47<04:24,  2.21s/it]\n",
      "Loss: 0.01111:  52%|█████▏    | 130/250 [04:47<04:24,  2.21s/it]\n",
      "Loss: 0.01200:  52%|█████▏    | 130/250 [04:47<04:24,  2.21s/it]\n",
      "Loss: 0.01366:  52%|█████▏    | 130/250 [04:47<04:24,  2.21s/it]\n",
      "Loss: 0.01058:  52%|█████▏    | 130/250 [04:48<04:24,  2.21s/it]\n",
      "Loss: 0.01045:  52%|█████▏    | 130/250 [04:48<04:24,  2.21s/it]\n",
      "Loss: 0.01210:  52%|█████▏    | 130/250 [04:48<04:24,  2.21s/it]\n",
      "Loss: 0.01474:  52%|█████▏    | 130/250 [04:48<04:24,  2.21s/it]\n",
      "Loss: 0.01168:  52%|█████▏    | 130/250 [04:48<04:24,  2.21s/it]\n",
      "Loss: 0.01122:  52%|█████▏    | 130/250 [04:48<04:24,  2.21s/it]\n",
      "Loss: 0.01373:  52%|█████▏    | 130/250 [04:48<04:24,  2.21s/it]\n",
      "Loss: 0.01218:  52%|█████▏    | 130/250 [04:48<04:24,  2.21s/it]\n",
      "Loss: 0.01265:  52%|█████▏    | 130/250 [04:48<04:24,  2.21s/it]\n",
      "Loss: 0.01293:  52%|█████▏    | 130/250 [04:48<04:24,  2.21s/it]\n",
      "Loss: 0.01245:  52%|█████▏    | 130/250 [04:49<04:24,  2.21s/it]\n",
      "Loss: 0.01245:  52%|█████▏    | 130/250 [04:49<04:24,  2.21s/it]\n",
      "Loss: 0.01155:  52%|█████▏    | 130/250 [04:49<04:24,  2.21s/it]\n",
      "Loss: 0.01278:  52%|█████▏    | 130/250 [04:49<04:24,  2.21s/it]\n",
      "Loss: 0.01136:  52%|█████▏    | 130/250 [04:49<04:24,  2.21s/it]\n",
      "Loss: 0.01190:  52%|█████▏    | 130/250 [04:49<04:24,  2.21s/it]\n",
      "Loss: 0.01184:  52%|█████▏    | 130/250 [04:49<04:24,  2.21s/it]\n",
      "Loss: 0.01193:  52%|█████▏    | 130/250 [04:49<04:24,  2.21s/it]\n",
      "Loss: 0.01193:  52%|█████▏    | 131/250 [04:49<04:22,  2.21s/it]\n",
      "Loss: 0.01122:  52%|█████▏    | 131/250 [04:49<04:22,  2.21s/it]\n",
      "Loss: 0.01178:  52%|█████▏    | 131/250 [04:50<04:22,  2.21s/it]\n",
      "Loss: 0.01170:  52%|█████▏    | 131/250 [04:50<04:22,  2.21s/it]\n",
      "Loss: 0.01342:  52%|█████▏    | 131/250 [04:50<04:22,  2.21s/it]\n",
      "Loss: 0.01290:  52%|█████▏    | 131/250 [04:50<04:22,  2.21s/it]\n",
      "Loss: 0.01223:  52%|█████▏    | 131/250 [04:50<04:22,  2.21s/it]\n",
      "Loss: 0.01354:  52%|█████▏    | 131/250 [04:50<04:22,  2.21s/it]\n",
      "Loss: 0.01148:  52%|█████▏    | 131/250 [04:50<04:22,  2.21s/it]\n",
      "Loss: 0.01381:  52%|█████▏    | 131/250 [04:50<04:22,  2.21s/it]\n",
      "Loss: 0.01232:  52%|█████▏    | 131/250 [04:50<04:22,  2.21s/it]\n",
      "Loss: 0.01125:  52%|█████▏    | 131/250 [04:50<04:22,  2.21s/it]\n",
      "Loss: 0.01089:  52%|█████▏    | 131/250 [04:51<04:22,  2.21s/it]\n",
      "Loss: 0.01291:  52%|█████▏    | 131/250 [04:51<04:22,  2.21s/it]\n",
      "Loss: 0.01232:  52%|█████▏    | 131/250 [04:51<04:22,  2.21s/it]\n",
      "Loss: 0.01257:  52%|█████▏    | 131/250 [04:51<04:22,  2.21s/it]\n",
      "Loss: 0.01170:  52%|█████▏    | 131/250 [04:51<04:22,  2.21s/it]\n",
      "Loss: 0.01266:  52%|█████▏    | 131/250 [04:51<04:22,  2.21s/it]\n",
      "Loss: 0.01173:  52%|█████▏    | 131/250 [04:51<04:22,  2.21s/it]\n",
      "Loss: 0.01093:  52%|█████▏    | 131/250 [04:51<04:22,  2.21s/it]\n",
      "Loss: 0.01179:  52%|█████▏    | 131/250 [04:51<04:22,  2.21s/it]\n",
      "Loss: 0.01224:  52%|█████▏    | 131/250 [04:51<04:22,  2.21s/it]\n",
      "Loss: 0.01224:  53%|█████▎    | 132/250 [04:51<04:20,  2.21s/it]\n",
      "Loss: 0.01174:  53%|█████▎    | 132/250 [04:52<04:20,  2.21s/it]\n",
      "Loss: 0.01178:  53%|█████▎    | 132/250 [04:52<04:20,  2.21s/it]\n",
      "Loss: 0.01329:  53%|█████▎    | 132/250 [04:52<04:20,  2.21s/it]\n",
      "Loss: 0.01043:  53%|█████▎    | 132/250 [04:52<04:20,  2.21s/it]\n",
      "Loss: 0.01029:  53%|█████▎    | 132/250 [04:52<04:20,  2.21s/it]\n",
      "Loss: 0.01204:  53%|█████▎    | 132/250 [04:52<04:20,  2.21s/it]\n",
      "Loss: 0.01303:  53%|█████▎    | 132/250 [04:52<04:20,  2.21s/it]\n",
      "Loss: 0.01156:  53%|█████▎    | 132/250 [04:52<04:20,  2.21s/it]\n",
      "Loss: 0.01374:  53%|█████▎    | 132/250 [04:52<04:20,  2.21s/it]\n",
      "Loss: 0.01226:  53%|█████▎    | 132/250 [04:53<04:20,  2.21s/it]\n",
      "Loss: 0.01136:  53%|█████▎    | 132/250 [04:53<04:20,  2.21s/it]\n",
      "Loss: 0.01124:  53%|█████▎    | 132/250 [04:53<04:20,  2.21s/it]\n",
      "Loss: 0.01200:  53%|█████▎    | 132/250 [04:53<04:20,  2.21s/it]\n",
      "Loss: 0.01314:  53%|█████▎    | 132/250 [04:53<04:20,  2.21s/it]\n",
      "Loss: 0.01242:  53%|█████▎    | 132/250 [04:53<04:20,  2.21s/it]\n",
      "Loss: 0.01107:  53%|█████▎    | 132/250 [04:53<04:20,  2.21s/it]\n",
      "Loss: 0.01276:  53%|█████▎    | 132/250 [04:53<04:20,  2.21s/it]\n",
      "Loss: 0.01163:  53%|█████▎    | 132/250 [04:53<04:20,  2.21s/it]\n",
      "Loss: 0.01289:  53%|█████▎    | 132/250 [04:54<04:20,  2.21s/it]\n",
      "Loss: 0.01171:  53%|█████▎    | 132/250 [04:54<04:20,  2.21s/it]\n",
      "Loss: 0.01207:  53%|█████▎    | 132/250 [04:54<04:20,  2.21s/it]\n",
      "Loss: 0.01207:  53%|█████▎    | 133/250 [04:54<04:17,  2.20s/it]\n",
      "Loss: 0.01130:  53%|█████▎    | 133/250 [04:54<04:17,  2.20s/it]\n",
      "Loss: 0.01235:  53%|█████▎    | 133/250 [04:54<04:17,  2.20s/it]\n",
      "Loss: 0.01356:  53%|█████▎    | 133/250 [04:54<04:17,  2.20s/it]\n",
      "Loss: 0.01033:  53%|█████▎    | 133/250 [04:54<04:17,  2.20s/it]\n",
      "Loss: 0.01010:  53%|█████▎    | 133/250 [04:54<04:17,  2.20s/it]\n",
      "Loss: 0.01209:  53%|█████▎    | 133/250 [04:54<04:17,  2.20s/it]\n",
      "Loss: 0.01001:  53%|█████▎    | 133/250 [04:54<04:17,  2.20s/it]\n",
      "Loss: 0.01515:  53%|█████▎    | 133/250 [04:55<04:17,  2.20s/it]\n",
      "Loss: 0.01047:  53%|█████▎    | 133/250 [04:55<04:17,  2.20s/it]\n",
      "Loss: 0.01333:  53%|█████▎    | 133/250 [04:55<04:17,  2.20s/it]\n",
      "Loss: 0.01364:  53%|█████▎    | 133/250 [04:55<04:17,  2.20s/it]\n",
      "Loss: 0.01157:  53%|█████▎    | 133/250 [04:55<04:17,  2.20s/it]\n",
      "Loss: 0.01265:  53%|█████▎    | 133/250 [04:55<04:17,  2.20s/it]\n",
      "Loss: 0.01345:  53%|█████▎    | 133/250 [04:55<04:17,  2.20s/it]\n",
      "Loss: 0.01123:  53%|█████▎    | 133/250 [04:55<04:17,  2.20s/it]\n",
      "Loss: 0.01087:  53%|█████▎    | 133/250 [04:55<04:17,  2.20s/it]\n",
      "Loss: 0.01105:  53%|█████▎    | 133/250 [04:55<04:17,  2.20s/it]\n",
      "Loss: 0.01307:  53%|█████▎    | 133/250 [04:56<04:17,  2.20s/it]\n",
      "Loss: 0.01326:  53%|█████▎    | 133/250 [04:56<04:17,  2.20s/it]\n",
      "Loss: 0.01199:  53%|█████▎    | 133/250 [04:56<04:17,  2.20s/it]\n",
      "Loss: 0.01236:  53%|█████▎    | 133/250 [04:56<04:17,  2.20s/it]\n",
      "Loss: 0.01236:  54%|█████▎    | 134/250 [04:56<04:15,  2.20s/it]\n",
      "Loss: 0.01133:  54%|█████▎    | 134/250 [04:56<04:15,  2.20s/it]\n",
      "Loss: 0.01188:  54%|█████▎    | 134/250 [04:56<04:15,  2.20s/it]\n",
      "Loss: 0.01153:  54%|█████▎    | 134/250 [04:56<04:15,  2.20s/it]\n",
      "Loss: 0.01328:  54%|█████▎    | 134/250 [04:56<04:15,  2.20s/it]\n",
      "Loss: 0.01275:  54%|█████▎    | 134/250 [04:56<04:15,  2.20s/it]\n",
      "Loss: 0.01192:  54%|█████▎    | 134/250 [04:57<04:15,  2.20s/it]\n",
      "Loss: 0.01297:  54%|█████▎    | 134/250 [04:57<04:15,  2.20s/it]\n",
      "Loss: 0.01142:  54%|█████▎    | 134/250 [04:57<04:15,  2.20s/it]\n",
      "Loss: 0.01382:  54%|█████▎    | 134/250 [04:57<04:15,  2.20s/it]\n",
      "Loss: 0.01256:  54%|█████▎    | 134/250 [04:57<04:15,  2.20s/it]\n",
      "Loss: 0.01205:  54%|█████▎    | 134/250 [04:57<04:15,  2.20s/it]\n",
      "Loss: 0.01271:  54%|█████▎    | 134/250 [04:57<04:15,  2.20s/it]\n",
      "Loss: 0.01097:  54%|█████▎    | 134/250 [04:57<04:15,  2.20s/it]\n",
      "Loss: 0.01035:  54%|█████▎    | 134/250 [04:57<04:15,  2.20s/it]\n",
      "Loss: 0.01068:  54%|█████▎    | 134/250 [04:58<04:15,  2.20s/it]\n",
      "Loss: 0.01060:  54%|█████▎    | 134/250 [04:58<04:15,  2.20s/it]\n",
      "Loss: 0.01211:  54%|█████▎    | 134/250 [04:58<04:15,  2.20s/it]\n",
      "Loss: 0.01221:  54%|█████▎    | 134/250 [04:58<04:15,  2.20s/it]\n",
      "Loss: 0.01076:  54%|█████▎    | 134/250 [04:58<04:15,  2.20s/it]\n",
      "Loss: 0.01204:  54%|█████▎    | 134/250 [04:58<04:15,  2.20s/it]\n",
      "Loss: 0.01204:  54%|█████▍    | 135/250 [04:58<04:12,  2.20s/it]\n",
      "Loss: 0.01197:  54%|█████▍    | 135/250 [04:58<04:12,  2.20s/it]\n",
      "Loss: 0.01171:  54%|█████▍    | 135/250 [04:58<04:12,  2.20s/it]\n",
      "Loss: 0.01199:  54%|█████▍    | 135/250 [04:58<04:12,  2.20s/it]\n",
      "Loss: 0.01310:  54%|█████▍    | 135/250 [04:58<04:12,  2.20s/it]\n",
      "Loss: 0.01312:  54%|█████▍    | 135/250 [04:59<04:12,  2.20s/it]\n",
      "Loss: 0.01206:  54%|█████▍    | 135/250 [04:59<04:12,  2.20s/it]\n",
      "Loss: 0.01040:  54%|█████▍    | 135/250 [04:59<04:12,  2.20s/it]\n",
      "Loss: 0.01243:  54%|█████▍    | 135/250 [04:59<04:12,  2.20s/it]\n",
      "Loss: 0.01103:  54%|█████▍    | 135/250 [04:59<04:12,  2.20s/it]\n",
      "Loss: 0.01307:  54%|█████▍    | 135/250 [04:59<04:12,  2.20s/it]\n",
      "Loss: 0.01206:  54%|█████▍    | 135/250 [04:59<04:12,  2.20s/it]\n",
      "Loss: 0.01261:  54%|█████▍    | 135/250 [04:59<04:12,  2.20s/it]\n",
      "Loss: 0.01296:  54%|█████▍    | 135/250 [04:59<04:12,  2.20s/it]\n",
      "Loss: 0.01211:  54%|█████▍    | 135/250 [05:00<04:12,  2.20s/it]\n",
      "Loss: 0.01260:  54%|█████▍    | 135/250 [05:00<04:12,  2.20s/it]\n",
      "Loss: 0.01024:  54%|█████▍    | 135/250 [05:00<04:12,  2.20s/it]\n",
      "Loss: 0.01223:  54%|█████▍    | 135/250 [05:00<04:12,  2.20s/it]\n",
      "Loss: 0.01176:  54%|█████▍    | 135/250 [05:00<04:12,  2.20s/it]\n",
      "Loss: 0.01057:  54%|█████▍    | 135/250 [05:00<04:12,  2.20s/it]\n",
      "Loss: 0.01155:  54%|█████▍    | 135/250 [05:00<04:12,  2.20s/it]\n",
      "Loss: 0.01208:  54%|█████▍    | 135/250 [05:00<04:12,  2.20s/it]\n",
      "Loss: 0.01208:  54%|█████▍    | 136/250 [05:00<04:10,  2.20s/it]\n",
      "Loss: 0.01104:  54%|█████▍    | 136/250 [05:00<04:10,  2.20s/it]\n",
      "Loss: 0.01167:  54%|█████▍    | 136/250 [05:00<04:10,  2.20s/it]\n",
      "Loss: 0.01341:  54%|█████▍    | 136/250 [05:01<04:10,  2.20s/it]\n",
      "Loss: 0.01018:  54%|█████▍    | 136/250 [05:01<04:10,  2.20s/it]\n",
      "Loss: 0.00984:  54%|█████▍    | 136/250 [05:01<04:10,  2.20s/it]\n",
      "Loss: 0.01226:  54%|█████▍    | 136/250 [05:01<04:10,  2.20s/it]\n",
      "Loss: 0.00953:  54%|█████▍    | 136/250 [05:01<04:10,  2.20s/it]\n",
      "Loss: 0.01552:  54%|█████▍    | 136/250 [05:01<04:10,  2.20s/it]\n",
      "Loss: 0.01017:  54%|█████▍    | 136/250 [05:01<04:10,  2.20s/it]\n",
      "Loss: 0.01331:  54%|█████▍    | 136/250 [05:01<04:10,  2.20s/it]\n",
      "Loss: 0.01365:  54%|█████▍    | 136/250 [05:01<04:10,  2.20s/it]\n",
      "Loss: 0.01177:  54%|█████▍    | 136/250 [05:02<04:10,  2.20s/it]\n",
      "Loss: 0.01258:  54%|█████▍    | 136/250 [05:02<04:10,  2.20s/it]\n",
      "Loss: 0.01334:  54%|█████▍    | 136/250 [05:02<04:10,  2.20s/it]\n",
      "Loss: 0.01273:  54%|█████▍    | 136/250 [05:02<04:10,  2.20s/it]\n",
      "Loss: 0.01045:  54%|█████▍    | 136/250 [05:02<04:10,  2.20s/it]\n",
      "Loss: 0.01166:  54%|█████▍    | 136/250 [05:02<04:10,  2.20s/it]\n",
      "Loss: 0.01199:  54%|█████▍    | 136/250 [05:02<04:10,  2.20s/it]\n",
      "Loss: 0.01424:  54%|█████▍    | 136/250 [05:02<04:10,  2.20s/it]\n",
      "Loss: 0.01137:  54%|█████▍    | 136/250 [05:02<04:10,  2.20s/it]\n",
      "Loss: 0.01206:  54%|█████▍    | 136/250 [05:02<04:10,  2.20s/it]\n",
      "Loss: 0.01206:  55%|█████▍    | 137/250 [05:02<04:09,  2.21s/it]\n",
      "Loss: 0.01142:  55%|█████▍    | 137/250 [05:03<04:09,  2.21s/it]\n",
      "Loss: 0.01204:  55%|█████▍    | 137/250 [05:03<04:09,  2.21s/it]\n",
      "Loss: 0.01330:  55%|█████▍    | 137/250 [05:03<04:09,  2.21s/it]\n",
      "Loss: 0.01025:  55%|█████▍    | 137/250 [05:03<04:09,  2.21s/it]\n",
      "Loss: 0.01007:  55%|█████▍    | 137/250 [05:03<04:09,  2.21s/it]\n",
      "Loss: 0.01228:  55%|█████▍    | 137/250 [05:03<04:09,  2.21s/it]\n",
      "Loss: 0.00968:  55%|█████▍    | 137/250 [05:03<04:09,  2.21s/it]\n",
      "Loss: 0.01497:  55%|█████▍    | 137/250 [05:03<04:09,  2.21s/it]\n",
      "Loss: 0.01026:  55%|█████▍    | 137/250 [05:03<04:09,  2.21s/it]\n",
      "Loss: 0.01392:  55%|█████▍    | 137/250 [05:04<04:09,  2.21s/it]\n",
      "Loss: 0.01389:  55%|█████▍    | 137/250 [05:04<04:09,  2.21s/it]\n",
      "Loss: 0.01117:  55%|█████▍    | 137/250 [05:04<04:09,  2.21s/it]\n",
      "Loss: 0.01173:  55%|█████▍    | 137/250 [05:04<04:09,  2.21s/it]\n",
      "Loss: 0.01035:  55%|█████▍    | 137/250 [05:04<04:09,  2.21s/it]\n",
      "Loss: 0.01123:  55%|█████▍    | 137/250 [05:04<04:09,  2.21s/it]\n",
      "Loss: 0.01323:  55%|█████▍    | 137/250 [05:04<04:09,  2.21s/it]\n",
      "Loss: 0.01079:  55%|█████▍    | 137/250 [05:04<04:09,  2.21s/it]\n",
      "Loss: 0.01325:  55%|█████▍    | 137/250 [05:04<04:09,  2.21s/it]\n",
      "Loss: 0.01252:  55%|█████▍    | 137/250 [05:05<04:09,  2.21s/it]\n",
      "Loss: 0.01144:  55%|█████▍    | 137/250 [05:05<04:09,  2.21s/it]\n",
      "Loss: 0.01199:  55%|█████▍    | 137/250 [05:05<04:09,  2.21s/it]\n",
      "Loss: 0.01199:  55%|█████▌    | 138/250 [05:05<04:07,  2.21s/it]\n",
      "Loss: 0.01115:  55%|█████▌    | 138/250 [05:05<04:07,  2.21s/it]\n",
      "Loss: 0.01187:  55%|█████▌    | 138/250 [05:05<04:07,  2.21s/it]\n",
      "Loss: 0.01329:  55%|█████▌    | 138/250 [05:05<04:07,  2.21s/it]\n",
      "Loss: 0.01021:  55%|█████▌    | 138/250 [05:05<04:07,  2.21s/it]\n",
      "Loss: 0.01001:  55%|█████▌    | 138/250 [05:05<04:07,  2.21s/it]\n",
      "Loss: 0.01246:  55%|█████▌    | 138/250 [05:05<04:07,  2.21s/it]\n",
      "Loss: 0.00977:  55%|█████▌    | 138/250 [05:05<04:07,  2.21s/it]\n",
      "Loss: 0.01536:  55%|█████▌    | 138/250 [05:06<04:07,  2.21s/it]\n",
      "Loss: 0.01019:  55%|█████▌    | 138/250 [05:06<04:07,  2.21s/it]\n",
      "Loss: 0.01361:  55%|█████▌    | 138/250 [05:06<04:07,  2.21s/it]\n",
      "Loss: 0.01440:  55%|█████▌    | 138/250 [05:06<04:07,  2.21s/it]\n",
      "Loss: 0.01142:  55%|█████▌    | 138/250 [05:06<04:07,  2.21s/it]\n",
      "Loss: 0.01274:  55%|█████▌    | 138/250 [05:06<04:07,  2.21s/it]\n",
      "Loss: 0.01330:  55%|█████▌    | 138/250 [05:06<04:07,  2.21s/it]\n",
      "Loss: 0.01164:  55%|█████▌    | 138/250 [05:06<04:07,  2.21s/it]\n",
      "Loss: 0.01063:  55%|█████▌    | 138/250 [05:06<04:07,  2.21s/it]\n",
      "Loss: 0.01164:  55%|█████▌    | 138/250 [05:07<04:07,  2.21s/it]\n",
      "Loss: 0.01173:  55%|█████▌    | 138/250 [05:07<04:07,  2.21s/it]\n",
      "Loss: 0.01430:  55%|█████▌    | 138/250 [05:07<04:07,  2.21s/it]\n",
      "Loss: 0.01162:  55%|█████▌    | 138/250 [05:07<04:07,  2.21s/it]\n",
      "Loss: 0.01226:  55%|█████▌    | 138/250 [05:07<04:07,  2.21s/it]\n",
      "Loss: 0.01226:  56%|█████▌    | 139/250 [05:07<04:05,  2.21s/it]\n",
      "Loss: 0.01094:  56%|█████▌    | 139/250 [05:07<04:05,  2.21s/it]\n",
      "Loss: 0.01260:  56%|█████▌    | 139/250 [05:07<04:05,  2.21s/it]\n",
      "Loss: 0.01345:  56%|█████▌    | 139/250 [05:07<04:05,  2.21s/it]\n",
      "Loss: 0.01030:  56%|█████▌    | 139/250 [05:07<04:05,  2.21s/it]\n",
      "Loss: 0.00975:  56%|█████▌    | 139/250 [05:07<04:05,  2.21s/it]\n",
      "Loss: 0.01238:  56%|█████▌    | 139/250 [05:08<04:05,  2.21s/it]\n",
      "Loss: 0.00990:  56%|█████▌    | 139/250 [05:08<04:05,  2.21s/it]\n",
      "Loss: 0.01496:  56%|█████▌    | 139/250 [05:08<04:05,  2.21s/it]\n",
      "Loss: 0.01053:  56%|█████▌    | 139/250 [05:08<04:05,  2.21s/it]\n",
      "Loss: 0.01209:  56%|█████▌    | 139/250 [05:08<04:05,  2.21s/it]\n",
      "Loss: 0.01083:  56%|█████▌    | 139/250 [05:08<04:05,  2.21s/it]\n",
      "Loss: 0.01165:  56%|█████▌    | 139/250 [05:08<04:05,  2.21s/it]\n",
      "Loss: 0.01204:  56%|█████▌    | 139/250 [05:08<04:05,  2.21s/it]\n",
      "Loss: 0.01030:  56%|█████▌    | 139/250 [05:08<04:05,  2.21s/it]\n",
      "Loss: 0.01222:  56%|█████▌    | 139/250 [05:09<04:05,  2.21s/it]\n",
      "Loss: 0.01158:  56%|█████▌    | 139/250 [05:09<04:05,  2.21s/it]\n",
      "Loss: 0.01254:  56%|█████▌    | 139/250 [05:09<04:05,  2.21s/it]\n",
      "Loss: 0.01121:  56%|█████▌    | 139/250 [05:09<04:05,  2.21s/it]\n",
      "Loss: 0.01159:  56%|█████▌    | 139/250 [05:09<04:05,  2.21s/it]\n",
      "Loss: 0.01141:  56%|█████▌    | 139/250 [05:09<04:05,  2.21s/it]\n",
      "Loss: 0.01190:  56%|█████▌    | 139/250 [05:09<04:05,  2.21s/it]\n",
      "Loss: 0.01190:  56%|█████▌    | 140/250 [05:09<04:03,  2.21s/it]\n",
      "Loss: 0.01197:  56%|█████▌    | 140/250 [05:09<04:03,  2.21s/it]\n",
      "Loss: 0.01073:  56%|█████▌    | 140/250 [05:09<04:03,  2.21s/it]\n",
      "Loss: 0.01160:  56%|█████▌    | 140/250 [05:09<04:03,  2.21s/it]\n",
      "Loss: 0.01137:  56%|█████▌    | 140/250 [05:10<04:03,  2.21s/it]\n",
      "Loss: 0.00958:  56%|█████▌    | 140/250 [05:10<04:03,  2.21s/it]\n",
      "Loss: 0.01119:  56%|█████▌    | 140/250 [05:10<04:03,  2.21s/it]\n",
      "Loss: 0.01215:  56%|█████▌    | 140/250 [05:10<04:03,  2.21s/it]\n",
      "Loss: 0.01253:  56%|█████▌    | 140/250 [05:10<04:03,  2.21s/it]\n",
      "Loss: 0.01085:  56%|█████▌    | 140/250 [05:10<04:03,  2.21s/it]\n",
      "Loss: 0.01090:  56%|█████▌    | 140/250 [05:10<04:03,  2.21s/it]\n",
      "Loss: 0.01209:  56%|█████▌    | 140/250 [05:10<04:03,  2.21s/it]\n",
      "Loss: 0.01127:  56%|█████▌    | 140/250 [05:10<04:03,  2.21s/it]\n",
      "Loss: 0.01091:  56%|█████▌    | 140/250 [05:11<04:03,  2.21s/it]\n",
      "Loss: 0.00996:  56%|█████▌    | 140/250 [05:11<04:03,  2.21s/it]\n",
      "Loss: 0.01070:  56%|█████▌    | 140/250 [05:11<04:03,  2.21s/it]\n",
      "Loss: 0.01051:  56%|█████▌    | 140/250 [05:11<04:03,  2.21s/it]\n",
      "Loss: 0.01196:  56%|█████▌    | 140/250 [05:11<04:03,  2.21s/it]\n",
      "Loss: 0.01170:  56%|█████▌    | 140/250 [05:11<04:03,  2.21s/it]\n",
      "Loss: 0.01078:  56%|█████▌    | 140/250 [05:11<04:03,  2.21s/it]\n",
      "Loss: 0.01219:  56%|█████▌    | 140/250 [05:11<04:03,  2.21s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 57.08it/s]\u001b[A\n",
      "Loss: 0.01219:  56%|█████▋    | 141/250 [05:11<03:59,  2.20s/it]\n",
      "Loss: 0.01074:  56%|█████▋    | 141/250 [05:11<03:59,  2.20s/it]\n",
      "Loss: 0.01175:  56%|█████▋    | 141/250 [05:12<03:59,  2.20s/it]\n",
      "Loss: 0.01308:  56%|█████▋    | 141/250 [05:12<03:59,  2.20s/it]\n",
      "Loss: 0.01321:  56%|█████▋    | 141/250 [05:12<03:59,  2.20s/it]\n",
      "Loss: 0.01332:  56%|█████▋    | 141/250 [05:12<03:59,  2.20s/it]\n",
      "Loss: 0.01190:  56%|█████▋    | 141/250 [05:12<03:59,  2.20s/it]\n",
      "Loss: 0.01296:  56%|█████▋    | 141/250 [05:12<03:59,  2.20s/it]\n",
      "Loss: 0.01161:  56%|█████▋    | 141/250 [05:12<03:59,  2.20s/it]\n",
      "Loss: 0.01350:  56%|█████▋    | 141/250 [05:12<03:59,  2.20s/it]\n",
      "Loss: 0.01203:  56%|█████▋    | 141/250 [05:12<03:59,  2.20s/it]\n",
      "Loss: 0.01081:  56%|█████▋    | 141/250 [05:12<03:59,  2.20s/it]\n",
      "Loss: 0.01136:  56%|█████▋    | 141/250 [05:13<03:59,  2.20s/it]\n",
      "Loss: 0.01197:  56%|█████▋    | 141/250 [05:13<03:59,  2.20s/it]\n",
      "Loss: 0.01057:  56%|█████▋    | 141/250 [05:13<03:59,  2.20s/it]\n",
      "Loss: 0.01094:  56%|█████▋    | 141/250 [05:13<03:59,  2.20s/it]\n",
      "Loss: 0.01320:  56%|█████▋    | 141/250 [05:13<03:59,  2.20s/it]\n",
      "Loss: 0.01069:  56%|█████▋    | 141/250 [05:13<03:59,  2.20s/it]\n",
      "Loss: 0.01319:  56%|█████▋    | 141/250 [05:13<03:59,  2.20s/it]\n",
      "Loss: 0.01284:  56%|█████▋    | 141/250 [05:13<03:59,  2.20s/it]\n",
      "Loss: 0.01160:  56%|█████▋    | 141/250 [05:13<03:59,  2.20s/it]\n",
      "Loss: 0.01188:  56%|█████▋    | 141/250 [05:13<03:59,  2.20s/it]\n",
      "Loss: 0.01188:  57%|█████▋    | 142/250 [05:14<03:57,  2.20s/it]\n",
      "Loss: 0.01081:  57%|█████▋    | 142/250 [05:14<03:57,  2.20s/it]\n",
      "Loss: 0.01172:  57%|█████▋    | 142/250 [05:14<03:57,  2.20s/it]\n",
      "Loss: 0.01377:  57%|█████▋    | 142/250 [05:14<03:57,  2.20s/it]\n",
      "Loss: 0.01002:  57%|█████▋    | 142/250 [05:14<03:57,  2.20s/it]\n",
      "Loss: 0.01006:  57%|█████▋    | 142/250 [05:14<03:57,  2.20s/it]\n",
      "Loss: 0.01201:  57%|█████▋    | 142/250 [05:14<03:57,  2.20s/it]\n",
      "Loss: 0.00933:  57%|█████▋    | 142/250 [05:14<03:57,  2.20s/it]\n",
      "Loss: 0.01496:  57%|█████▋    | 142/250 [05:14<03:57,  2.20s/it]\n",
      "Loss: 0.01058:  57%|█████▋    | 142/250 [05:14<03:57,  2.20s/it]\n",
      "Loss: 0.01363:  57%|█████▋    | 142/250 [05:15<03:57,  2.20s/it]\n",
      "Loss: 0.01308:  57%|█████▋    | 142/250 [05:15<03:57,  2.20s/it]\n",
      "Loss: 0.01149:  57%|█████▋    | 142/250 [05:15<03:57,  2.20s/it]\n",
      "Loss: 0.01241:  57%|█████▋    | 142/250 [05:15<03:57,  2.20s/it]\n",
      "Loss: 0.01330:  57%|█████▋    | 142/250 [05:15<03:57,  2.20s/it]\n",
      "Loss: 0.01140:  57%|█████▋    | 142/250 [05:15<03:57,  2.20s/it]\n",
      "Loss: 0.01034:  57%|█████▋    | 142/250 [05:15<03:57,  2.20s/it]\n",
      "Loss: 0.01167:  57%|█████▋    | 142/250 [05:15<03:57,  2.20s/it]\n",
      "Loss: 0.01195:  57%|█████▋    | 142/250 [05:15<03:57,  2.20s/it]\n",
      "Loss: 0.01447:  57%|█████▋    | 142/250 [05:16<03:57,  2.20s/it]\n",
      "Loss: 0.01142:  57%|█████▋    | 142/250 [05:16<03:57,  2.20s/it]\n",
      "Loss: 0.01210:  57%|█████▋    | 142/250 [05:16<03:57,  2.20s/it]\n",
      "Loss: 0.01210:  57%|█████▋    | 143/250 [05:16<03:56,  2.21s/it]\n",
      "Loss: 0.01175:  57%|█████▋    | 143/250 [05:16<03:56,  2.21s/it]\n",
      "Loss: 0.01133:  57%|█████▋    | 143/250 [05:16<03:56,  2.21s/it]\n",
      "Loss: 0.01166:  57%|█████▋    | 143/250 [05:16<03:56,  2.21s/it]\n",
      "Loss: 0.01315:  57%|█████▋    | 143/250 [05:16<03:56,  2.21s/it]\n",
      "Loss: 0.01301:  57%|█████▋    | 143/250 [05:16<03:56,  2.21s/it]\n",
      "Loss: 0.01019:  57%|█████▋    | 143/250 [05:16<03:56,  2.21s/it]\n",
      "Loss: 0.01070:  57%|█████▋    | 143/250 [05:16<03:56,  2.21s/it]\n",
      "Loss: 0.01255:  57%|█████▋    | 143/250 [05:17<03:56,  2.21s/it]\n",
      "Loss: 0.01046:  57%|█████▋    | 143/250 [05:17<03:56,  2.21s/it]\n",
      "Loss: 0.01091:  57%|█████▋    | 143/250 [05:17<03:56,  2.21s/it]\n",
      "Loss: 0.01210:  57%|█████▋    | 143/250 [05:17<03:56,  2.21s/it]\n",
      "Loss: 0.01140:  57%|█████▋    | 143/250 [05:17<03:56,  2.21s/it]\n",
      "Loss: 0.01116:  57%|█████▋    | 143/250 [05:17<03:56,  2.21s/it]\n",
      "Loss: 0.00988:  57%|█████▋    | 143/250 [05:17<03:56,  2.21s/it]\n",
      "Loss: 0.01046:  57%|█████▋    | 143/250 [05:17<03:56,  2.21s/it]\n",
      "Loss: 0.01033:  57%|█████▋    | 143/250 [05:17<03:56,  2.21s/it]\n",
      "Loss: 0.01183:  57%|█████▋    | 143/250 [05:18<03:56,  2.21s/it]\n",
      "Loss: 0.01107:  57%|█████▋    | 143/250 [05:18<03:56,  2.21s/it]\n",
      "Loss: 0.01271:  57%|█████▋    | 143/250 [05:18<03:56,  2.21s/it]\n",
      "Loss: 0.01160:  57%|█████▋    | 143/250 [05:18<03:56,  2.21s/it]\n",
      "Loss: 0.01185:  57%|█████▋    | 143/250 [05:18<03:56,  2.21s/it]\n",
      "Loss: 0.01185:  58%|█████▊    | 144/250 [05:18<03:53,  2.20s/it]\n",
      "Loss: 0.01128:  58%|█████▊    | 144/250 [05:18<03:53,  2.20s/it]\n",
      "Loss: 0.01185:  58%|█████▊    | 144/250 [05:18<03:53,  2.20s/it]\n",
      "Loss: 0.01146:  58%|█████▊    | 144/250 [05:18<03:53,  2.20s/it]\n",
      "Loss: 0.01301:  58%|█████▊    | 144/250 [05:18<03:53,  2.20s/it]\n",
      "Loss: 0.01308:  58%|█████▊    | 144/250 [05:18<03:53,  2.20s/it]\n",
      "Loss: 0.01202:  58%|█████▊    | 144/250 [05:19<03:53,  2.20s/it]\n",
      "Loss: 0.01268:  58%|█████▊    | 144/250 [05:19<03:53,  2.20s/it]\n",
      "Loss: 0.01177:  58%|█████▊    | 144/250 [05:19<03:53,  2.20s/it]\n",
      "Loss: 0.01346:  58%|█████▊    | 144/250 [05:19<03:53,  2.20s/it]\n",
      "Loss: 0.01230:  58%|█████▊    | 144/250 [05:19<03:53,  2.20s/it]\n",
      "Loss: 0.01089:  58%|█████▊    | 144/250 [05:19<03:53,  2.20s/it]\n",
      "Loss: 0.01091:  58%|█████▊    | 144/250 [05:19<03:53,  2.20s/it]\n",
      "Loss: 0.01143:  58%|█████▊    | 144/250 [05:19<03:53,  2.20s/it]\n",
      "Loss: 0.01019:  58%|█████▊    | 144/250 [05:19<03:53,  2.20s/it]\n",
      "Loss: 0.01095:  58%|█████▊    | 144/250 [05:20<03:53,  2.20s/it]\n",
      "Loss: 0.01294:  58%|█████▊    | 144/250 [05:20<03:53,  2.20s/it]\n",
      "Loss: 0.01104:  58%|█████▊    | 144/250 [05:20<03:53,  2.20s/it]\n",
      "Loss: 0.01309:  58%|█████▊    | 144/250 [05:20<03:53,  2.20s/it]\n",
      "Loss: 0.01298:  58%|█████▊    | 144/250 [05:20<03:53,  2.20s/it]\n",
      "Loss: 0.01131:  58%|█████▊    | 144/250 [05:20<03:53,  2.20s/it]\n",
      "Loss: 0.01234:  58%|█████▊    | 144/250 [05:20<03:53,  2.20s/it]\n",
      "Loss: 0.01234:  58%|█████▊    | 145/250 [05:20<03:52,  2.21s/it]\n",
      "Loss: 0.01104:  58%|█████▊    | 145/250 [05:20<03:52,  2.21s/it]\n",
      "Loss: 0.01173:  58%|█████▊    | 145/250 [05:20<03:52,  2.21s/it]\n",
      "Loss: 0.01340:  58%|█████▊    | 145/250 [05:20<03:52,  2.21s/it]\n",
      "Loss: 0.01031:  58%|█████▊    | 145/250 [05:21<03:52,  2.21s/it]\n",
      "Loss: 0.01001:  58%|█████▊    | 145/250 [05:21<03:52,  2.21s/it]\n",
      "Loss: 0.01187:  58%|█████▊    | 145/250 [05:21<03:52,  2.21s/it]\n",
      "Loss: 0.00978:  58%|█████▊    | 145/250 [05:21<03:52,  2.21s/it]\n",
      "Loss: 0.01455:  58%|█████▊    | 145/250 [05:21<03:52,  2.21s/it]\n",
      "Loss: 0.00999:  58%|█████▊    | 145/250 [05:21<03:52,  2.21s/it]\n",
      "Loss: 0.01304:  58%|█████▊    | 145/250 [05:21<03:52,  2.21s/it]\n",
      "Loss: 0.01303:  58%|█████▊    | 145/250 [05:21<03:52,  2.21s/it]\n",
      "Loss: 0.01104:  58%|█████▊    | 145/250 [05:21<03:52,  2.21s/it]\n",
      "Loss: 0.01201:  58%|█████▊    | 145/250 [05:22<03:52,  2.21s/it]\n",
      "Loss: 0.01335:  58%|█████▊    | 145/250 [05:22<03:52,  2.21s/it]\n",
      "Loss: 0.01116:  58%|█████▊    | 145/250 [05:22<03:52,  2.21s/it]\n",
      "Loss: 0.01036:  58%|█████▊    | 145/250 [05:22<03:52,  2.21s/it]\n",
      "Loss: 0.01180:  58%|█████▊    | 145/250 [05:22<03:52,  2.21s/it]\n",
      "Loss: 0.01199:  58%|█████▊    | 145/250 [05:22<03:52,  2.21s/it]\n",
      "Loss: 0.01425:  58%|█████▊    | 145/250 [05:22<03:52,  2.21s/it]\n",
      "Loss: 0.01106:  58%|█████▊    | 145/250 [05:22<03:52,  2.21s/it]\n",
      "Loss: 0.01191:  58%|█████▊    | 145/250 [05:22<03:52,  2.21s/it]\n",
      "Loss: 0.01191:  58%|█████▊    | 146/250 [05:22<03:50,  2.21s/it]\n",
      "Loss: 0.01218:  58%|█████▊    | 146/250 [05:22<03:50,  2.21s/it]\n",
      "Loss: 0.01128:  58%|█████▊    | 146/250 [05:23<03:50,  2.21s/it]\n",
      "Loss: 0.01162:  58%|█████▊    | 146/250 [05:23<03:50,  2.21s/it]\n",
      "Loss: 0.01303:  58%|█████▊    | 146/250 [05:23<03:50,  2.21s/it]\n",
      "Loss: 0.00993:  58%|█████▊    | 146/250 [05:23<03:50,  2.21s/it]\n",
      "Loss: 0.01163:  58%|█████▊    | 146/250 [05:23<03:50,  2.21s/it]\n",
      "Loss: 0.00933:  58%|█████▊    | 146/250 [05:23<03:50,  2.21s/it]\n",
      "Loss: 0.01462:  58%|█████▊    | 146/250 [05:23<03:50,  2.21s/it]\n",
      "Loss: 0.01018:  58%|█████▊    | 146/250 [05:23<03:50,  2.21s/it]\n",
      "Loss: 0.01349:  58%|█████▊    | 146/250 [05:23<03:50,  2.21s/it]\n",
      "Loss: 0.01442:  58%|█████▊    | 146/250 [05:24<03:50,  2.21s/it]\n",
      "Loss: 0.01110:  58%|█████▊    | 146/250 [05:24<03:50,  2.21s/it]\n",
      "Loss: 0.01211:  58%|█████▊    | 146/250 [05:24<03:50,  2.21s/it]\n",
      "Loss: 0.01335:  58%|█████▊    | 146/250 [05:24<03:50,  2.21s/it]\n",
      "Loss: 0.01150:  58%|█████▊    | 146/250 [05:24<03:50,  2.21s/it]\n",
      "Loss: 0.01045:  58%|█████▊    | 146/250 [05:24<03:50,  2.21s/it]\n",
      "Loss: 0.01156:  58%|█████▊    | 146/250 [05:24<03:50,  2.21s/it]\n",
      "Loss: 0.01214:  58%|█████▊    | 146/250 [05:24<03:50,  2.21s/it]\n",
      "Loss: 0.01421:  58%|█████▊    | 146/250 [05:24<03:50,  2.21s/it]\n",
      "Loss: 0.01132:  58%|█████▊    | 146/250 [05:24<03:50,  2.21s/it]\n",
      "Loss: 0.01208:  58%|█████▊    | 146/250 [05:25<03:50,  2.21s/it]\n",
      "Loss: 0.01208:  59%|█████▉    | 147/250 [05:25<03:48,  2.22s/it]\n",
      "Loss: 0.01086:  59%|█████▉    | 147/250 [05:25<03:48,  2.22s/it]\n",
      "Loss: 0.01190:  59%|█████▉    | 147/250 [05:25<03:48,  2.22s/it]\n",
      "Loss: 0.01143:  59%|█████▉    | 147/250 [05:25<03:48,  2.22s/it]\n",
      "Loss: 0.01316:  59%|█████▉    | 147/250 [05:25<03:48,  2.22s/it]\n",
      "Loss: 0.01267:  59%|█████▉    | 147/250 [05:25<03:48,  2.22s/it]\n",
      "Loss: 0.01224:  59%|█████▉    | 147/250 [05:25<03:48,  2.22s/it]\n",
      "Loss: 0.01255:  59%|█████▉    | 147/250 [05:25<03:48,  2.22s/it]\n",
      "Loss: 0.01133:  59%|█████▉    | 147/250 [05:25<03:48,  2.22s/it]\n",
      "Loss: 0.01327:  59%|█████▉    | 147/250 [05:26<03:48,  2.22s/it]\n",
      "Loss: 0.01203:  59%|█████▉    | 147/250 [05:26<03:48,  2.22s/it]\n",
      "Loss: 0.01081:  59%|█████▉    | 147/250 [05:26<03:48,  2.22s/it]\n",
      "Loss: 0.01244:  59%|█████▉    | 147/250 [05:26<03:48,  2.22s/it]\n",
      "Loss: 0.01261:  59%|█████▉    | 147/250 [05:26<03:48,  2.22s/it]\n",
      "Loss: 0.01194:  59%|█████▉    | 147/250 [05:26<03:48,  2.22s/it]\n",
      "Loss: 0.01218:  59%|█████▉    | 147/250 [05:26<03:48,  2.22s/it]\n",
      "Loss: 0.01110:  59%|█████▉    | 147/250 [05:26<03:48,  2.22s/it]\n",
      "Loss: 0.01262:  59%|█████▉    | 147/250 [05:26<03:48,  2.22s/it]\n",
      "Loss: 0.01112:  59%|█████▉    | 147/250 [05:27<03:48,  2.22s/it]\n",
      "Loss: 0.01180:  59%|█████▉    | 147/250 [05:27<03:48,  2.22s/it]\n",
      "Loss: 0.01175:  59%|█████▉    | 147/250 [05:27<03:48,  2.22s/it]\n",
      "Loss: 0.01196:  59%|█████▉    | 147/250 [05:27<03:48,  2.22s/it]\n",
      "Loss: 0.01196:  59%|█████▉    | 148/250 [05:27<03:45,  2.22s/it]\n",
      "Loss: 0.01077:  59%|█████▉    | 148/250 [05:27<03:45,  2.22s/it]\n",
      "Loss: 0.01165:  59%|█████▉    | 148/250 [05:27<03:45,  2.22s/it]\n",
      "Loss: 0.01340:  59%|█████▉    | 148/250 [05:27<03:45,  2.22s/it]\n",
      "Loss: 0.01018:  59%|█████▉    | 148/250 [05:27<03:45,  2.22s/it]\n",
      "Loss: 0.00989:  59%|█████▉    | 148/250 [05:27<03:45,  2.22s/it]\n",
      "Loss: 0.01179:  59%|█████▉    | 148/250 [05:27<03:45,  2.22s/it]\n",
      "Loss: 0.00984:  59%|█████▉    | 148/250 [05:28<03:45,  2.22s/it]\n",
      "Loss: 0.01452:  59%|█████▉    | 148/250 [05:28<03:45,  2.22s/it]\n",
      "Loss: 0.01027:  59%|█████▉    | 148/250 [05:28<03:45,  2.22s/it]\n",
      "Loss: 0.01296:  59%|█████▉    | 148/250 [05:28<03:45,  2.22s/it]\n",
      "Loss: 0.01352:  59%|█████▉    | 148/250 [05:28<03:45,  2.22s/it]\n",
      "Loss: 0.01110:  59%|█████▉    | 148/250 [05:28<03:45,  2.22s/it]\n",
      "Loss: 0.01218:  59%|█████▉    | 148/250 [05:28<03:45,  2.22s/it]\n",
      "Loss: 0.01333:  59%|█████▉    | 148/250 [05:28<03:45,  2.22s/it]\n",
      "Loss: 0.01096:  59%|█████▉    | 148/250 [05:28<03:45,  2.22s/it]\n",
      "Loss: 0.01063:  59%|█████▉    | 148/250 [05:29<03:45,  2.22s/it]\n",
      "Loss: 0.01163:  59%|█████▉    | 148/250 [05:29<03:45,  2.22s/it]\n",
      "Loss: 0.01172:  59%|█████▉    | 148/250 [05:29<03:45,  2.22s/it]\n",
      "Loss: 0.01275:  59%|█████▉    | 148/250 [05:29<03:45,  2.22s/it]\n",
      "Loss: 0.01123:  59%|█████▉    | 148/250 [05:29<03:45,  2.22s/it]\n",
      "Loss: 0.01240:  59%|█████▉    | 148/250 [05:29<03:45,  2.22s/it]\n",
      "Loss: 0.01240:  60%|█████▉    | 149/250 [05:29<03:43,  2.21s/it]\n",
      "Loss: 0.01088:  60%|█████▉    | 149/250 [05:29<03:43,  2.21s/it]\n",
      "Loss: 0.01160:  60%|█████▉    | 149/250 [05:29<03:43,  2.21s/it]\n",
      "Loss: 0.01306:  60%|█████▉    | 149/250 [05:29<03:43,  2.21s/it]\n",
      "Loss: 0.00978:  60%|█████▉    | 149/250 [05:29<03:43,  2.21s/it]\n",
      "Loss: 0.01009:  60%|█████▉    | 149/250 [05:30<03:43,  2.21s/it]\n",
      "Loss: 0.01176:  60%|█████▉    | 149/250 [05:30<03:43,  2.21s/it]\n",
      "Loss: 0.00960:  60%|█████▉    | 149/250 [05:30<03:43,  2.21s/it]\n",
      "Loss: 0.01453:  60%|█████▉    | 149/250 [05:30<03:43,  2.21s/it]\n",
      "Loss: 0.01017:  60%|█████▉    | 149/250 [05:30<03:43,  2.21s/it]\n",
      "Loss: 0.01295:  60%|█████▉    | 149/250 [05:30<03:43,  2.21s/it]\n",
      "Loss: 0.01331:  60%|█████▉    | 149/250 [05:30<03:43,  2.21s/it]\n",
      "Loss: 0.01139:  60%|█████▉    | 149/250 [05:30<03:43,  2.21s/it]\n",
      "Loss: 0.01202:  60%|█████▉    | 149/250 [05:30<03:43,  2.21s/it]\n",
      "Loss: 0.01324:  60%|█████▉    | 149/250 [05:30<03:43,  2.21s/it]\n",
      "Loss: 0.01107:  60%|█████▉    | 149/250 [05:31<03:43,  2.21s/it]\n",
      "Loss: 0.01271:  60%|█████▉    | 149/250 [05:31<03:43,  2.21s/it]\n",
      "Loss: 0.01053:  60%|█████▉    | 149/250 [05:31<03:43,  2.21s/it]\n",
      "Loss: 0.01302:  60%|█████▉    | 149/250 [05:31<03:43,  2.21s/it]\n",
      "Loss: 0.01296:  60%|█████▉    | 149/250 [05:31<03:43,  2.21s/it]\n",
      "Loss: 0.01148:  60%|█████▉    | 149/250 [05:31<03:43,  2.21s/it]\n",
      "Loss: 0.01182:  60%|█████▉    | 149/250 [05:31<03:43,  2.21s/it]\n",
      "Loss: 0.01182:  60%|██████    | 150/250 [05:31<03:41,  2.21s/it]\n",
      "Loss: 0.01096:  60%|██████    | 150/250 [05:31<03:41,  2.21s/it]\n",
      "Loss: 0.01148:  60%|██████    | 150/250 [05:31<03:41,  2.21s/it]\n",
      "Loss: 0.01322:  60%|██████    | 150/250 [05:32<03:41,  2.21s/it]\n",
      "Loss: 0.01030:  60%|██████    | 150/250 [05:32<03:41,  2.21s/it]\n",
      "Loss: 0.00991:  60%|██████    | 150/250 [05:32<03:41,  2.21s/it]\n",
      "Loss: 0.01175:  60%|██████    | 150/250 [05:32<03:41,  2.21s/it]\n",
      "Loss: 0.01074:  60%|██████    | 150/250 [05:32<03:41,  2.21s/it]\n",
      "Loss: 0.01495:  60%|██████    | 150/250 [05:32<03:41,  2.21s/it]\n",
      "Loss: 0.01012:  60%|██████    | 150/250 [05:32<03:41,  2.21s/it]\n",
      "Loss: 0.01215:  60%|██████    | 150/250 [05:32<03:41,  2.21s/it]\n",
      "Loss: 0.01064:  60%|██████    | 150/250 [05:32<03:41,  2.21s/it]\n",
      "Loss: 0.01077:  60%|██████    | 150/250 [05:32<03:41,  2.21s/it]\n",
      "Loss: 0.01170:  60%|██████    | 150/250 [05:33<03:41,  2.21s/it]\n",
      "Loss: 0.01020:  60%|██████    | 150/250 [05:33<03:41,  2.21s/it]\n",
      "Loss: 0.01080:  60%|██████    | 150/250 [05:33<03:41,  2.21s/it]\n",
      "Loss: 0.01295:  60%|██████    | 150/250 [05:33<03:41,  2.21s/it]\n",
      "Loss: 0.01032:  60%|██████    | 150/250 [05:33<03:41,  2.21s/it]\n",
      "Loss: 0.01283:  60%|██████    | 150/250 [05:33<03:41,  2.21s/it]\n",
      "Loss: 0.01242:  60%|██████    | 150/250 [05:33<03:41,  2.21s/it]\n",
      "Loss: 0.01156:  60%|██████    | 150/250 [05:33<03:41,  2.21s/it]\n",
      "Loss: 0.01206:  60%|██████    | 150/250 [05:33<03:41,  2.21s/it]\n",
      "Loss: 0.01206:  60%|██████    | 151/250 [05:33<03:38,  2.21s/it]\n",
      "Loss: 0.01102:  60%|██████    | 151/250 [05:34<03:38,  2.21s/it]\n",
      "Loss: 0.01151:  60%|██████    | 151/250 [05:34<03:38,  2.21s/it]\n",
      "Loss: 0.01312:  60%|██████    | 151/250 [05:34<03:38,  2.21s/it]\n",
      "Loss: 0.00995:  60%|██████    | 151/250 [05:34<03:38,  2.21s/it]\n",
      "Loss: 0.00990:  60%|██████    | 151/250 [05:34<03:38,  2.21s/it]\n",
      "Loss: 0.01158:  60%|██████    | 151/250 [05:34<03:38,  2.21s/it]\n",
      "Loss: 0.00949:  60%|██████    | 151/250 [05:34<03:38,  2.21s/it]\n",
      "Loss: 0.01470:  60%|██████    | 151/250 [05:34<03:38,  2.21s/it]\n",
      "Loss: 0.01006:  60%|██████    | 151/250 [05:34<03:38,  2.21s/it]\n",
      "Loss: 0.01306:  60%|██████    | 151/250 [05:34<03:38,  2.21s/it]\n",
      "Loss: 0.01343:  60%|██████    | 151/250 [05:35<03:38,  2.21s/it]\n",
      "Loss: 0.01138:  60%|██████    | 151/250 [05:35<03:38,  2.21s/it]\n",
      "Loss: 0.01225:  60%|██████    | 151/250 [05:35<03:38,  2.21s/it]\n",
      "Loss: 0.01303:  60%|██████    | 151/250 [05:35<03:38,  2.21s/it]\n",
      "Loss: 0.01105:  60%|██████    | 151/250 [05:35<03:38,  2.21s/it]\n",
      "Loss: 0.01084:  60%|██████    | 151/250 [05:35<03:38,  2.21s/it]\n",
      "Loss: 0.01188:  60%|██████    | 151/250 [05:35<03:38,  2.21s/it]\n",
      "Loss: 0.01160:  60%|██████    | 151/250 [05:35<03:38,  2.21s/it]\n",
      "Loss: 0.01378:  60%|██████    | 151/250 [05:35<03:38,  2.21s/it]\n",
      "Loss: 0.01161:  60%|██████    | 151/250 [05:36<03:38,  2.21s/it]\n",
      "Loss: 0.01210:  60%|██████    | 151/250 [05:36<03:38,  2.21s/it]\n",
      "Loss: 0.01210:  61%|██████    | 152/250 [05:36<03:36,  2.21s/it]\n",
      "Loss: 0.01068:  61%|██████    | 152/250 [05:36<03:36,  2.21s/it]\n",
      "Loss: 0.01158:  61%|██████    | 152/250 [05:36<03:36,  2.21s/it]\n",
      "Loss: 0.01316:  61%|██████    | 152/250 [05:36<03:36,  2.21s/it]\n",
      "Loss: 0.01278:  61%|██████    | 152/250 [05:36<03:36,  2.21s/it]\n",
      "Loss: 0.01272:  61%|██████    | 152/250 [05:36<03:36,  2.21s/it]\n",
      "Loss: 0.01184:  61%|██████    | 152/250 [05:36<03:36,  2.21s/it]\n",
      "Loss: 0.01257:  61%|██████    | 152/250 [05:36<03:36,  2.21s/it]\n",
      "Loss: 0.01106:  61%|██████    | 152/250 [05:36<03:36,  2.21s/it]\n",
      "Loss: 0.01339:  61%|██████    | 152/250 [05:37<03:36,  2.21s/it]\n",
      "Loss: 0.01180:  61%|██████    | 152/250 [05:37<03:36,  2.21s/it]\n",
      "Loss: 0.01095:  61%|██████    | 152/250 [05:37<03:36,  2.21s/it]\n",
      "Loss: 0.01092:  61%|██████    | 152/250 [05:37<03:36,  2.21s/it]\n",
      "Loss: 0.01139:  61%|██████    | 152/250 [05:37<03:36,  2.21s/it]\n",
      "Loss: 0.01038:  61%|██████    | 152/250 [05:37<03:36,  2.21s/it]\n",
      "Loss: 0.01041:  61%|██████    | 152/250 [05:37<03:36,  2.21s/it]\n",
      "Loss: 0.01287:  61%|██████    | 152/250 [05:37<03:36,  2.21s/it]\n",
      "Loss: 0.01027:  61%|██████    | 152/250 [05:37<03:36,  2.21s/it]\n",
      "Loss: 0.01271:  61%|██████    | 152/250 [05:38<03:36,  2.21s/it]\n",
      "Loss: 0.01242:  61%|██████    | 152/250 [05:38<03:36,  2.21s/it]\n",
      "Loss: 0.01120:  61%|██████    | 152/250 [05:38<03:36,  2.21s/it]\n",
      "Loss: 0.01161:  61%|██████    | 152/250 [05:38<03:36,  2.21s/it]\n",
      "Loss: 0.01161:  61%|██████    | 153/250 [05:38<03:35,  2.22s/it]\n",
      "Loss: 0.01082:  61%|██████    | 153/250 [05:38<03:35,  2.22s/it]\n",
      "Loss: 0.01155:  61%|██████    | 153/250 [05:38<03:35,  2.22s/it]\n",
      "Loss: 0.01318:  61%|██████    | 153/250 [05:38<03:35,  2.22s/it]\n",
      "Loss: 0.01286:  61%|██████    | 153/250 [05:38<03:35,  2.22s/it]\n",
      "Loss: 0.01271:  61%|██████    | 153/250 [05:38<03:35,  2.22s/it]\n",
      "Loss: 0.01218:  61%|██████    | 153/250 [05:39<03:35,  2.22s/it]\n",
      "Loss: 0.01319:  61%|██████    | 153/250 [05:39<03:35,  2.22s/it]\n",
      "Loss: 0.01193:  61%|██████    | 153/250 [05:39<03:35,  2.22s/it]\n",
      "Loss: 0.01096:  61%|██████    | 153/250 [05:39<03:35,  2.22s/it]\n",
      "Loss: 0.01300:  61%|██████    | 153/250 [05:39<03:35,  2.22s/it]\n",
      "Loss: 0.01184:  61%|██████    | 153/250 [05:39<03:35,  2.22s/it]\n",
      "Loss: 0.01210:  61%|██████    | 153/250 [05:39<03:35,  2.22s/it]\n",
      "Loss: 0.01260:  61%|██████    | 153/250 [05:39<03:35,  2.22s/it]\n",
      "Loss: 0.01210:  61%|██████    | 153/250 [05:39<03:35,  2.22s/it]\n",
      "Loss: 0.01180:  61%|██████    | 153/250 [05:39<03:35,  2.22s/it]\n",
      "Loss: 0.01151:  61%|██████    | 153/250 [05:40<03:35,  2.22s/it]\n",
      "Loss: 0.01222:  61%|██████    | 153/250 [05:40<03:35,  2.22s/it]\n",
      "Loss: 0.01099:  61%|██████    | 153/250 [05:40<03:35,  2.22s/it]\n",
      "Loss: 0.01158:  61%|██████    | 153/250 [05:40<03:35,  2.22s/it]\n",
      "Loss: 0.01172:  61%|██████    | 153/250 [05:40<03:35,  2.22s/it]\n",
      "Loss: 0.01195:  61%|██████    | 153/250 [05:40<03:35,  2.22s/it]\n",
      "Loss: 0.01195:  62%|██████▏   | 154/250 [05:40<03:32,  2.22s/it]\n",
      "Loss: 0.01088:  62%|██████▏   | 154/250 [05:40<03:32,  2.22s/it]\n",
      "Loss: 0.01154:  62%|██████▏   | 154/250 [05:40<03:32,  2.22s/it]\n",
      "Loss: 0.01322:  62%|██████▏   | 154/250 [05:40<03:32,  2.22s/it]\n",
      "Loss: 0.01301:  62%|██████▏   | 154/250 [05:41<03:32,  2.22s/it]\n",
      "Loss: 0.01238:  62%|██████▏   | 154/250 [05:41<03:32,  2.22s/it]\n",
      "Loss: 0.01170:  62%|██████▏   | 154/250 [05:41<03:32,  2.22s/it]\n",
      "Loss: 0.01256:  62%|██████▏   | 154/250 [05:41<03:32,  2.22s/it]\n",
      "Loss: 0.01125:  62%|██████▏   | 154/250 [05:41<03:32,  2.22s/it]\n",
      "Loss: 0.01006:  62%|██████▏   | 154/250 [05:41<03:32,  2.22s/it]\n",
      "Loss: 0.01325:  62%|██████▏   | 154/250 [05:41<03:32,  2.22s/it]\n",
      "Loss: 0.01082:  62%|██████▏   | 154/250 [05:41<03:32,  2.22s/it]\n",
      "Loss: 0.01104:  62%|██████▏   | 154/250 [05:41<03:32,  2.22s/it]\n",
      "Loss: 0.01130:  62%|██████▏   | 154/250 [05:42<03:32,  2.22s/it]\n",
      "Loss: 0.01036:  62%|██████▏   | 154/250 [05:42<03:32,  2.22s/it]\n",
      "Loss: 0.01080:  62%|██████▏   | 154/250 [05:42<03:32,  2.22s/it]\n",
      "Loss: 0.01255:  62%|██████▏   | 154/250 [05:42<03:32,  2.22s/it]\n",
      "Loss: 0.01076:  62%|██████▏   | 154/250 [05:42<03:32,  2.22s/it]\n",
      "Loss: 0.01262:  62%|██████▏   | 154/250 [05:42<03:32,  2.22s/it]\n",
      "Loss: 0.01200:  62%|██████▏   | 154/250 [05:42<03:32,  2.22s/it]\n",
      "Loss: 0.01167:  62%|██████▏   | 154/250 [05:42<03:32,  2.22s/it]\n",
      "Loss: 0.01172:  62%|██████▏   | 154/250 [05:42<03:32,  2.22s/it]\n",
      "Loss: 0.01172:  62%|██████▏   | 155/250 [05:42<03:32,  2.24s/it]\n",
      "Loss: 0.01177:  62%|██████▏   | 155/250 [05:43<03:32,  2.24s/it]\n",
      "Loss: 0.01133:  62%|██████▏   | 155/250 [05:43<03:32,  2.24s/it]\n",
      "Loss: 0.01136:  62%|██████▏   | 155/250 [05:43<03:32,  2.24s/it]\n",
      "Loss: 0.01286:  62%|██████▏   | 155/250 [05:43<03:32,  2.24s/it]\n",
      "Loss: 0.01264:  62%|██████▏   | 155/250 [05:43<03:32,  2.24s/it]\n",
      "Loss: 0.01176:  62%|██████▏   | 155/250 [05:43<03:32,  2.24s/it]\n",
      "Loss: 0.01224:  62%|██████▏   | 155/250 [05:43<03:32,  2.24s/it]\n",
      "Loss: 0.01104:  62%|██████▏   | 155/250 [05:43<03:32,  2.24s/it]\n",
      "Loss: 0.01324:  62%|██████▏   | 155/250 [05:43<03:32,  2.24s/it]\n",
      "Loss: 0.01216:  62%|██████▏   | 155/250 [05:43<03:32,  2.24s/it]\n",
      "Loss: 0.01075:  62%|██████▏   | 155/250 [05:44<03:32,  2.24s/it]\n",
      "Loss: 0.01097:  62%|██████▏   | 155/250 [05:44<03:32,  2.24s/it]\n",
      "Loss: 0.01119:  62%|██████▏   | 155/250 [05:44<03:32,  2.24s/it]\n",
      "Loss: 0.01016:  62%|██████▏   | 155/250 [05:44<03:32,  2.24s/it]\n",
      "Loss: 0.01057:  62%|██████▏   | 155/250 [05:44<03:32,  2.24s/it]\n",
      "Loss: 0.01268:  62%|██████▏   | 155/250 [05:44<03:32,  2.24s/it]\n",
      "Loss: 0.01059:  62%|██████▏   | 155/250 [05:44<03:32,  2.24s/it]\n",
      "Loss: 0.01273:  62%|██████▏   | 155/250 [05:44<03:32,  2.24s/it]\n",
      "Loss: 0.01272:  62%|██████▏   | 155/250 [05:44<03:32,  2.24s/it]\n",
      "Loss: 0.01147:  62%|██████▏   | 155/250 [05:45<03:32,  2.24s/it]\n",
      "Loss: 0.01209:  62%|██████▏   | 155/250 [05:45<03:32,  2.24s/it]\n",
      "Loss: 0.01209:  62%|██████▏   | 156/250 [05:45<03:29,  2.23s/it]\n",
      "Loss: 0.01121:  62%|██████▏   | 156/250 [05:45<03:29,  2.23s/it]\n",
      "Loss: 0.01238:  62%|██████▏   | 156/250 [05:45<03:29,  2.23s/it]\n",
      "Loss: 0.01337:  62%|██████▏   | 156/250 [05:45<03:29,  2.23s/it]\n",
      "Loss: 0.00982:  62%|██████▏   | 156/250 [05:45<03:29,  2.23s/it]\n",
      "Loss: 0.00989:  62%|██████▏   | 156/250 [05:45<03:29,  2.23s/it]\n",
      "Loss: 0.01171:  62%|██████▏   | 156/250 [05:45<03:29,  2.23s/it]\n",
      "Loss: 0.00959:  62%|██████▏   | 156/250 [05:45<03:29,  2.23s/it]\n",
      "Loss: 0.01441:  62%|██████▏   | 156/250 [05:45<03:29,  2.23s/it]\n",
      "Loss: 0.00992:  62%|██████▏   | 156/250 [05:46<03:29,  2.23s/it]\n",
      "Loss: 0.01282:  62%|██████▏   | 156/250 [05:46<03:29,  2.23s/it]\n",
      "Loss: 0.01327:  62%|██████▏   | 156/250 [05:46<03:29,  2.23s/it]\n",
      "Loss: 0.01139:  62%|██████▏   | 156/250 [05:46<03:29,  2.23s/it]\n",
      "Loss: 0.01219:  62%|██████▏   | 156/250 [05:46<03:29,  2.23s/it]\n",
      "Loss: 0.01296:  62%|██████▏   | 156/250 [05:46<03:29,  2.23s/it]\n",
      "Loss: 0.01128:  62%|██████▏   | 156/250 [05:46<03:29,  2.23s/it]\n",
      "Loss: 0.01023:  62%|██████▏   | 156/250 [05:46<03:29,  2.23s/it]\n",
      "Loss: 0.01170:  62%|██████▏   | 156/250 [05:46<03:29,  2.23s/it]\n",
      "Loss: 0.01149:  62%|██████▏   | 156/250 [05:46<03:29,  2.23s/it]\n",
      "Loss: 0.01410:  62%|██████▏   | 156/250 [05:47<03:29,  2.23s/it]\n",
      "Loss: 0.01166:  62%|██████▏   | 156/250 [05:47<03:29,  2.23s/it]\n",
      "Loss: 0.01166:  62%|██████▏   | 156/250 [05:47<03:29,  2.23s/it]\n",
      "Loss: 0.01166:  63%|██████▎   | 157/250 [05:47<03:26,  2.23s/it]\n",
      "Loss: 0.01060:  63%|██████▎   | 157/250 [05:47<03:26,  2.23s/it]\n",
      "Loss: 0.01153:  63%|██████▎   | 157/250 [05:47<03:26,  2.23s/it]\n",
      "Loss: 0.01308:  63%|██████▎   | 157/250 [05:47<03:26,  2.23s/it]\n",
      "Loss: 0.00976:  63%|██████▎   | 157/250 [05:47<03:26,  2.23s/it]\n",
      "Loss: 0.00978:  63%|██████▎   | 157/250 [05:47<03:26,  2.23s/it]\n",
      "Loss: 0.01201:  63%|██████▎   | 157/250 [05:47<03:26,  2.23s/it]\n",
      "Loss: 0.00940:  63%|██████▎   | 157/250 [05:48<03:26,  2.23s/it]\n",
      "Loss: 0.01430:  63%|██████▎   | 157/250 [05:48<03:26,  2.23s/it]\n",
      "Loss: 0.01011:  63%|██████▎   | 157/250 [05:48<03:26,  2.23s/it]\n",
      "Loss: 0.01314:  63%|██████▎   | 157/250 [05:48<03:26,  2.23s/it]\n",
      "Loss: 0.01300:  63%|██████▎   | 157/250 [05:48<03:26,  2.23s/it]\n",
      "Loss: 0.01118:  63%|██████▎   | 157/250 [05:48<03:26,  2.23s/it]\n",
      "Loss: 0.01185:  63%|██████▎   | 157/250 [05:48<03:26,  2.23s/it]\n",
      "Loss: 0.01303:  63%|██████▎   | 157/250 [05:48<03:26,  2.23s/it]\n",
      "Loss: 0.01132:  63%|██████▎   | 157/250 [05:48<03:26,  2.23s/it]\n",
      "Loss: 0.01070:  63%|██████▎   | 157/250 [05:48<03:26,  2.23s/it]\n",
      "Loss: 0.01162:  63%|██████▎   | 157/250 [05:49<03:26,  2.23s/it]\n",
      "Loss: 0.01157:  63%|██████▎   | 157/250 [05:49<03:26,  2.23s/it]\n",
      "Loss: 0.01377:  63%|██████▎   | 157/250 [05:49<03:26,  2.23s/it]\n",
      "Loss: 0.01132:  63%|██████▎   | 157/250 [05:49<03:26,  2.23s/it]\n",
      "Loss: 0.01184:  63%|██████▎   | 157/250 [05:49<03:26,  2.23s/it]\n",
      "Loss: 0.01184:  63%|██████▎   | 158/250 [05:49<03:23,  2.22s/it]\n",
      "Loss: 0.01077:  63%|██████▎   | 158/250 [05:49<03:23,  2.22s/it]\n",
      "Loss: 0.01142:  63%|██████▎   | 158/250 [05:49<03:23,  2.22s/it]\n",
      "Loss: 0.01313:  63%|██████▎   | 158/250 [05:49<03:23,  2.22s/it]\n",
      "Loss: 0.00965:  63%|██████▎   | 158/250 [05:49<03:23,  2.22s/it]\n",
      "Loss: 0.00969:  63%|██████▎   | 158/250 [05:50<03:23,  2.22s/it]\n",
      "Loss: 0.01151:  63%|██████▎   | 158/250 [05:50<03:23,  2.22s/it]\n",
      "Loss: 0.00917:  63%|██████▎   | 158/250 [05:50<03:23,  2.22s/it]\n",
      "Loss: 0.01433:  63%|██████▎   | 158/250 [05:50<03:23,  2.22s/it]\n",
      "Loss: 0.00982:  63%|██████▎   | 158/250 [05:50<03:23,  2.22s/it]\n",
      "Loss: 0.01283:  63%|██████▎   | 158/250 [05:50<03:23,  2.22s/it]\n",
      "Loss: 0.01292:  63%|██████▎   | 158/250 [05:50<03:23,  2.22s/it]\n",
      "Loss: 0.01101:  63%|██████▎   | 158/250 [05:50<03:23,  2.22s/it]\n",
      "Loss: 0.01225:  63%|██████▎   | 158/250 [05:50<03:23,  2.22s/it]\n",
      "Loss: 0.01007:  63%|██████▎   | 158/250 [05:51<03:23,  2.22s/it]\n",
      "Loss: 0.01038:  63%|██████▎   | 158/250 [05:51<03:23,  2.22s/it]\n",
      "Loss: 0.01226:  63%|██████▎   | 158/250 [05:51<03:23,  2.22s/it]\n",
      "Loss: 0.01203:  63%|██████▎   | 158/250 [05:51<03:23,  2.22s/it]\n",
      "Loss: 0.01054:  63%|██████▎   | 158/250 [05:51<03:23,  2.22s/it]\n",
      "Loss: 0.01120:  63%|██████▎   | 158/250 [05:51<03:23,  2.22s/it]\n",
      "Loss: 0.01135:  63%|██████▎   | 158/250 [05:51<03:23,  2.22s/it]\n",
      "Loss: 0.01159:  63%|██████▎   | 158/250 [05:51<03:23,  2.22s/it]\n",
      "Loss: 0.01159:  64%|██████▎   | 159/250 [05:51<03:21,  2.21s/it]\n",
      "Loss: 0.01076:  64%|██████▎   | 159/250 [05:51<03:21,  2.21s/it]\n",
      "Loss: 0.01181:  64%|██████▎   | 159/250 [05:51<03:21,  2.21s/it]\n",
      "Loss: 0.01331:  64%|██████▎   | 159/250 [05:52<03:21,  2.21s/it]\n",
      "Loss: 0.00969:  64%|██████▎   | 159/250 [05:52<03:21,  2.21s/it]\n",
      "Loss: 0.00977:  64%|██████▎   | 159/250 [05:52<03:21,  2.21s/it]\n",
      "Loss: 0.01180:  64%|██████▎   | 159/250 [05:52<03:21,  2.21s/it]\n",
      "Loss: 0.00946:  64%|██████▎   | 159/250 [05:52<03:21,  2.21s/it]\n",
      "Loss: 0.01127:  64%|██████▎   | 159/250 [05:52<03:21,  2.21s/it]\n",
      "Loss: 0.01380:  64%|██████▎   | 159/250 [05:52<03:21,  2.21s/it]\n",
      "Loss: 0.01238:  64%|██████▎   | 159/250 [05:52<03:21,  2.21s/it]\n",
      "Loss: 0.01214:  64%|██████▎   | 159/250 [05:52<03:21,  2.21s/it]\n",
      "Loss: 0.01062:  64%|██████▎   | 159/250 [05:52<03:21,  2.21s/it]\n",
      "Loss: 0.01048:  64%|██████▎   | 159/250 [05:53<03:21,  2.21s/it]\n",
      "Loss: 0.01060:  64%|██████▎   | 159/250 [05:53<03:21,  2.21s/it]\n",
      "Loss: 0.01174:  64%|██████▎   | 159/250 [05:53<03:21,  2.21s/it]\n",
      "Loss: 0.00996:  64%|██████▎   | 159/250 [05:53<03:21,  2.21s/it]\n",
      "Loss: 0.01168:  64%|██████▎   | 159/250 [05:53<03:21,  2.21s/it]\n",
      "Loss: 0.01143:  64%|██████▎   | 159/250 [05:53<03:21,  2.21s/it]\n",
      "Loss: 0.01024:  64%|██████▎   | 159/250 [05:53<03:21,  2.21s/it]\n",
      "Loss: 0.01110:  64%|██████▎   | 159/250 [05:53<03:21,  2.21s/it]\n",
      "Loss: 0.01153:  64%|██████▎   | 159/250 [05:53<03:21,  2.21s/it]\n",
      "Loss: 0.01153:  64%|██████▍   | 160/250 [05:53<03:18,  2.21s/it]\n",
      "Loss: 0.01067:  64%|██████▍   | 160/250 [05:54<03:18,  2.21s/it]\n",
      "Loss: 0.01152:  64%|██████▍   | 160/250 [05:54<03:18,  2.21s/it]\n",
      "Loss: 0.01296:  64%|██████▍   | 160/250 [05:54<03:18,  2.21s/it]\n",
      "Loss: 0.00993:  64%|██████▍   | 160/250 [05:54<03:18,  2.21s/it]\n",
      "Loss: 0.00968:  64%|██████▍   | 160/250 [05:54<03:18,  2.21s/it]\n",
      "Loss: 0.01171:  64%|██████▍   | 160/250 [05:54<03:18,  2.21s/it]\n",
      "Loss: 0.01215:  64%|██████▍   | 160/250 [05:54<03:18,  2.21s/it]\n",
      "Loss: 0.01182:  64%|██████▍   | 160/250 [05:54<03:18,  2.21s/it]\n",
      "Loss: 0.01071:  64%|██████▍   | 160/250 [05:54<03:18,  2.21s/it]\n",
      "Loss: 0.01289:  64%|██████▍   | 160/250 [05:54<03:18,  2.21s/it]\n",
      "Loss: 0.01187:  64%|██████▍   | 160/250 [05:55<03:18,  2.21s/it]\n",
      "Loss: 0.01200:  64%|██████▍   | 160/250 [05:55<03:18,  2.21s/it]\n",
      "Loss: 0.01277:  64%|██████▍   | 160/250 [05:55<03:18,  2.21s/it]\n",
      "Loss: 0.01174:  64%|██████▍   | 160/250 [05:55<03:18,  2.21s/it]\n",
      "Loss: 0.01212:  64%|██████▍   | 160/250 [05:55<03:18,  2.21s/it]\n",
      "Loss: 0.01142:  64%|██████▍   | 160/250 [05:55<03:18,  2.21s/it]\n",
      "Loss: 0.01210:  64%|██████▍   | 160/250 [05:55<03:18,  2.21s/it]\n",
      "Loss: 0.01094:  64%|██████▍   | 160/250 [05:55<03:18,  2.21s/it]\n",
      "Loss: 0.01145:  64%|██████▍   | 160/250 [05:55<03:18,  2.21s/it]\n",
      "Loss: 0.01150:  64%|██████▍   | 160/250 [05:56<03:18,  2.21s/it]\n",
      "Loss: 0.01168:  64%|██████▍   | 160/250 [05:56<03:18,  2.21s/it]\n",
      "Loss: 0.01168:  64%|██████▍   | 161/250 [05:56<03:16,  2.20s/it]\n",
      "Loss: 0.01067:  64%|██████▍   | 161/250 [05:56<03:16,  2.20s/it]\n",
      "Loss: 0.01130:  64%|██████▍   | 161/250 [05:56<03:16,  2.20s/it]\n",
      "Loss: 0.01278:  64%|██████▍   | 161/250 [05:56<03:16,  2.20s/it]\n",
      "Loss: 0.00980:  64%|██████▍   | 161/250 [05:56<03:16,  2.20s/it]\n",
      "Loss: 0.00993:  64%|██████▍   | 161/250 [05:56<03:16,  2.20s/it]\n",
      "Loss: 0.01177:  64%|██████▍   | 161/250 [05:56<03:16,  2.20s/it]\n",
      "Loss: 0.00937:  64%|██████▍   | 161/250 [05:56<03:16,  2.20s/it]\n",
      "Loss: 0.01418:  64%|██████▍   | 161/250 [05:56<03:16,  2.20s/it]\n",
      "Loss: 0.00980:  64%|██████▍   | 161/250 [05:57<03:16,  2.20s/it]\n",
      "Loss: 0.01155:  64%|██████▍   | 161/250 [05:57<03:16,  2.20s/it]\n",
      "Loss: 0.01115:  64%|██████▍   | 161/250 [05:57<03:16,  2.20s/it]\n",
      "Loss: 0.01064:  64%|██████▍   | 161/250 [05:57<03:16,  2.20s/it]\n",
      "Loss: 0.01140:  64%|██████▍   | 161/250 [05:57<03:16,  2.20s/it]\n",
      "Loss: 0.00997:  64%|██████▍   | 161/250 [05:57<03:16,  2.20s/it]\n",
      "Loss: 0.01031:  64%|██████▍   | 161/250 [05:57<03:16,  2.20s/it]\n",
      "Loss: 0.01255:  64%|██████▍   | 161/250 [05:57<03:16,  2.20s/it]\n",
      "Loss: 0.01036:  64%|██████▍   | 161/250 [05:57<03:16,  2.20s/it]\n",
      "Loss: 0.01249:  64%|██████▍   | 161/250 [05:58<03:16,  2.20s/it]\n",
      "Loss: 0.01215:  64%|██████▍   | 161/250 [05:58<03:16,  2.20s/it]\n",
      "Loss: 0.01105:  64%|██████▍   | 161/250 [05:58<03:16,  2.20s/it]\n",
      "Loss: 0.01219:  64%|██████▍   | 161/250 [05:58<03:16,  2.20s/it]\n",
      "Loss: 0.01219:  65%|██████▍   | 162/250 [05:58<03:14,  2.21s/it]\n",
      "Loss: 0.01060:  65%|██████▍   | 162/250 [05:58<03:14,  2.21s/it]\n",
      "Loss: 0.01121:  65%|██████▍   | 162/250 [05:58<03:14,  2.21s/it]\n",
      "Loss: 0.01285:  65%|██████▍   | 162/250 [05:58<03:14,  2.21s/it]\n",
      "Loss: 0.00966:  65%|██████▍   | 162/250 [05:58<03:14,  2.21s/it]\n",
      "Loss: 0.00964:  65%|██████▍   | 162/250 [05:58<03:14,  2.21s/it]\n",
      "Loss: 0.01170:  65%|██████▍   | 162/250 [05:58<03:14,  2.21s/it]\n",
      "Loss: 0.00914:  65%|██████▍   | 162/250 [05:59<03:14,  2.21s/it]\n",
      "Loss: 0.01437:  65%|██████▍   | 162/250 [05:59<03:14,  2.21s/it]\n",
      "Loss: 0.00981:  65%|██████▍   | 162/250 [05:59<03:14,  2.21s/it]\n",
      "Loss: 0.01265:  65%|██████▍   | 162/250 [05:59<03:14,  2.21s/it]\n",
      "Loss: 0.01297:  65%|██████▍   | 162/250 [05:59<03:14,  2.21s/it]\n",
      "Loss: 0.01100:  65%|██████▍   | 162/250 [05:59<03:14,  2.21s/it]\n",
      "Loss: 0.01183:  65%|██████▍   | 162/250 [05:59<03:14,  2.21s/it]\n",
      "Loss: 0.01283:  65%|██████▍   | 162/250 [05:59<03:14,  2.21s/it]\n",
      "Loss: 0.01049:  65%|██████▍   | 162/250 [05:59<03:14,  2.21s/it]\n",
      "Loss: 0.01033:  65%|██████▍   | 162/250 [05:59<03:14,  2.21s/it]\n",
      "Loss: 0.01157:  65%|██████▍   | 162/250 [06:00<03:14,  2.21s/it]\n",
      "Loss: 0.01145:  65%|██████▍   | 162/250 [06:00<03:14,  2.21s/it]\n",
      "Loss: 0.01379:  65%|██████▍   | 162/250 [06:00<03:14,  2.21s/it]\n",
      "Loss: 0.01110:  65%|██████▍   | 162/250 [06:00<03:14,  2.21s/it]\n",
      "Loss: 0.01206:  65%|██████▍   | 162/250 [06:00<03:14,  2.21s/it]\n",
      "Loss: 0.01206:  65%|██████▌   | 163/250 [06:00<03:11,  2.21s/it]\n",
      "Loss: 0.01061:  65%|██████▌   | 163/250 [06:00<03:11,  2.21s/it]\n",
      "Loss: 0.01132:  65%|██████▌   | 163/250 [06:00<03:11,  2.21s/it]\n",
      "Loss: 0.01275:  65%|██████▌   | 163/250 [06:00<03:11,  2.21s/it]\n",
      "Loss: 0.00956:  65%|██████▌   | 163/250 [06:00<03:11,  2.21s/it]\n",
      "Loss: 0.00953:  65%|██████▌   | 163/250 [06:01<03:11,  2.21s/it]\n",
      "Loss: 0.01178:  65%|██████▌   | 163/250 [06:01<03:11,  2.21s/it]\n",
      "Loss: 0.00990:  65%|██████▌   | 163/250 [06:01<03:11,  2.21s/it]\n",
      "Loss: 0.01095:  65%|██████▌   | 163/250 [06:01<03:11,  2.21s/it]\n",
      "Loss: 0.01309:  65%|██████▌   | 163/250 [06:01<03:11,  2.21s/it]\n",
      "Loss: 0.01163:  65%|██████▌   | 163/250 [06:01<03:11,  2.21s/it]\n",
      "Loss: 0.01184:  65%|██████▌   | 163/250 [06:01<03:11,  2.21s/it]\n",
      "Loss: 0.01187:  65%|██████▌   | 163/250 [06:01<03:11,  2.21s/it]\n",
      "Loss: 0.01230:  65%|██████▌   | 163/250 [06:01<03:11,  2.21s/it]\n",
      "Loss: 0.01162:  65%|██████▌   | 163/250 [06:02<03:11,  2.21s/it]\n",
      "Loss: 0.01158:  65%|██████▌   | 163/250 [06:02<03:11,  2.21s/it]\n",
      "Loss: 0.01084:  65%|██████▌   | 163/250 [06:02<03:11,  2.21s/it]\n",
      "Loss: 0.01203:  65%|██████▌   | 163/250 [06:02<03:11,  2.21s/it]\n",
      "Loss: 0.01089:  65%|██████▌   | 163/250 [06:02<03:11,  2.21s/it]\n",
      "Loss: 0.01160:  65%|██████▌   | 163/250 [06:02<03:11,  2.21s/it]\n",
      "Loss: 0.01137:  65%|██████▌   | 163/250 [06:02<03:11,  2.21s/it]\n",
      "Loss: 0.01179:  65%|██████▌   | 163/250 [06:02<03:11,  2.21s/it]\n",
      "Loss: 0.01179:  66%|██████▌   | 164/250 [06:02<03:10,  2.21s/it]\n",
      "Loss: 0.01159:  66%|██████▌   | 164/250 [06:02<03:10,  2.21s/it]\n",
      "Loss: 0.01103:  66%|██████▌   | 164/250 [06:02<03:10,  2.21s/it]\n",
      "Loss: 0.01131:  66%|██████▌   | 164/250 [06:03<03:10,  2.21s/it]\n",
      "Loss: 0.01255:  66%|██████▌   | 164/250 [06:03<03:10,  2.21s/it]\n",
      "Loss: 0.01254:  66%|██████▌   | 164/250 [06:03<03:10,  2.21s/it]\n",
      "Loss: 0.01197:  66%|██████▌   | 164/250 [06:03<03:10,  2.21s/it]\n",
      "Loss: 0.01027:  66%|██████▌   | 164/250 [06:03<03:10,  2.21s/it]\n",
      "Loss: 0.01148:  66%|██████▌   | 164/250 [06:03<03:10,  2.21s/it]\n",
      "Loss: 0.01067:  66%|██████▌   | 164/250 [06:03<03:10,  2.21s/it]\n",
      "Loss: 0.01242:  66%|██████▌   | 164/250 [06:03<03:10,  2.21s/it]\n",
      "Loss: 0.01148:  66%|██████▌   | 164/250 [06:03<03:10,  2.21s/it]\n",
      "Loss: 0.01198:  66%|██████▌   | 164/250 [06:04<03:10,  2.21s/it]\n",
      "Loss: 0.01247:  66%|██████▌   | 164/250 [06:04<03:10,  2.21s/it]\n",
      "Loss: 0.01187:  66%|██████▌   | 164/250 [06:04<03:10,  2.21s/it]\n",
      "Loss: 0.01186:  66%|██████▌   | 164/250 [06:04<03:10,  2.21s/it]\n",
      "Loss: 0.01105:  66%|██████▌   | 164/250 [06:04<03:10,  2.21s/it]\n",
      "Loss: 0.01163:  66%|██████▌   | 164/250 [06:04<03:10,  2.21s/it]\n",
      "Loss: 0.01101:  66%|██████▌   | 164/250 [06:04<03:10,  2.21s/it]\n",
      "Loss: 0.01035:  66%|██████▌   | 164/250 [06:04<03:10,  2.21s/it]\n",
      "Loss: 0.01117:  66%|██████▌   | 164/250 [06:04<03:10,  2.21s/it]\n",
      "Loss: 0.01218:  66%|██████▌   | 164/250 [06:04<03:10,  2.21s/it]\n",
      "Loss: 0.01218:  66%|██████▌   | 165/250 [06:04<03:07,  2.21s/it]\n",
      "Loss: 0.01050:  66%|██████▌   | 165/250 [06:05<03:07,  2.21s/it]\n",
      "Loss: 0.01162:  66%|██████▌   | 165/250 [06:05<03:07,  2.21s/it]\n",
      "Loss: 0.01293:  66%|██████▌   | 165/250 [06:05<03:07,  2.21s/it]\n",
      "Loss: 0.00957:  66%|██████▌   | 165/250 [06:05<03:07,  2.21s/it]\n",
      "Loss: 0.00985:  66%|██████▌   | 165/250 [06:05<03:07,  2.21s/it]\n",
      "Loss: 0.01173:  66%|██████▌   | 165/250 [06:05<03:07,  2.21s/it]\n",
      "Loss: 0.01024:  66%|██████▌   | 165/250 [06:05<03:07,  2.21s/it]\n",
      "Loss: 0.01448:  66%|██████▌   | 165/250 [06:05<03:07,  2.21s/it]\n",
      "Loss: 0.01297:  66%|██████▌   | 165/250 [06:05<03:07,  2.21s/it]\n",
      "Loss: 0.01166:  66%|██████▌   | 165/250 [06:06<03:07,  2.21s/it]\n",
      "Loss: 0.01067:  66%|██████▌   | 165/250 [06:06<03:07,  2.21s/it]\n",
      "Loss: 0.01229:  66%|██████▌   | 165/250 [06:06<03:07,  2.21s/it]\n",
      "Loss: 0.01250:  66%|██████▌   | 165/250 [06:06<03:07,  2.21s/it]\n",
      "Loss: 0.01173:  66%|██████▌   | 165/250 [06:06<03:07,  2.21s/it]\n",
      "Loss: 0.01191:  66%|██████▌   | 165/250 [06:06<03:07,  2.21s/it]\n",
      "Loss: 0.01119:  66%|██████▌   | 165/250 [06:06<03:07,  2.21s/it]\n",
      "Loss: 0.01207:  66%|██████▌   | 165/250 [06:06<03:07,  2.21s/it]\n",
      "Loss: 0.01077:  66%|██████▌   | 165/250 [06:06<03:07,  2.21s/it]\n",
      "Loss: 0.01116:  66%|██████▌   | 165/250 [06:06<03:07,  2.21s/it]\n",
      "Loss: 0.01145:  66%|██████▌   | 165/250 [06:07<03:07,  2.21s/it]\n",
      "Loss: 0.01144:  66%|██████▌   | 165/250 [06:07<03:07,  2.21s/it]\n",
      "Loss: 0.01144:  66%|██████▋   | 166/250 [06:07<03:05,  2.21s/it]\n",
      "Loss: 0.01146:  66%|██████▋   | 166/250 [06:07<03:05,  2.21s/it]\n",
      "Loss: 0.01112:  66%|██████▋   | 166/250 [06:07<03:05,  2.21s/it]\n",
      "Loss: 0.01101:  66%|██████▋   | 166/250 [06:07<03:05,  2.21s/it]\n",
      "Loss: 0.01079:  66%|██████▋   | 166/250 [06:07<03:05,  2.21s/it]\n",
      "Loss: 0.00924:  66%|██████▋   | 166/250 [06:07<03:05,  2.21s/it]\n",
      "Loss: 0.01008:  66%|██████▋   | 166/250 [06:07<03:05,  2.21s/it]\n",
      "Loss: 0.01015:  66%|██████▋   | 166/250 [06:07<03:05,  2.21s/it]\n",
      "Loss: 0.01164:  66%|██████▋   | 166/250 [06:08<03:05,  2.21s/it]\n",
      "Loss: 0.01039:  66%|██████▋   | 166/250 [06:08<03:05,  2.21s/it]\n",
      "Loss: 0.01221:  66%|██████▋   | 166/250 [06:08<03:05,  2.21s/it]\n",
      "Loss: 0.01177:  66%|██████▋   | 166/250 [06:08<03:05,  2.21s/it]\n",
      "Loss: 0.01195:  66%|██████▋   | 166/250 [06:08<03:05,  2.21s/it]\n",
      "Loss: 0.01221:  66%|██████▋   | 166/250 [06:08<03:05,  2.21s/it]\n",
      "Loss: 0.01156:  66%|██████▋   | 166/250 [06:08<03:05,  2.21s/it]\n",
      "Loss: 0.01178:  66%|██████▋   | 166/250 [06:08<03:05,  2.21s/it]\n",
      "Loss: 0.01099:  66%|██████▋   | 166/250 [06:08<03:05,  2.21s/it]\n",
      "Loss: 0.01227:  66%|██████▋   | 166/250 [06:08<03:05,  2.21s/it]\n",
      "Loss: 0.01099:  66%|██████▋   | 166/250 [06:09<03:05,  2.21s/it]\n",
      "Loss: 0.01143:  66%|██████▋   | 166/250 [06:09<03:05,  2.21s/it]\n",
      "Loss: 0.01120:  66%|██████▋   | 166/250 [06:09<03:05,  2.21s/it]\n",
      "Loss: 0.01172:  66%|██████▋   | 166/250 [06:09<03:05,  2.21s/it]\n",
      "Loss: 0.01172:  67%|██████▋   | 167/250 [06:09<03:02,  2.20s/it]\n",
      "Loss: 0.01052:  67%|██████▋   | 167/250 [06:09<03:02,  2.20s/it]\n",
      "Loss: 0.01132:  67%|██████▋   | 167/250 [06:09<03:02,  2.20s/it]\n",
      "Loss: 0.01284:  67%|██████▋   | 167/250 [06:09<03:02,  2.20s/it]\n",
      "Loss: 0.00959:  67%|██████▋   | 167/250 [06:09<03:02,  2.20s/it]\n",
      "Loss: 0.00931:  67%|██████▋   | 167/250 [06:09<03:02,  2.20s/it]\n",
      "Loss: 0.01152:  67%|██████▋   | 167/250 [06:09<03:02,  2.20s/it]\n",
      "Loss: 0.00922:  67%|██████▋   | 167/250 [06:10<03:02,  2.20s/it]\n",
      "Loss: 0.01422:  67%|██████▋   | 167/250 [06:10<03:02,  2.20s/it]\n",
      "Loss: 0.00957:  67%|██████▋   | 167/250 [06:10<03:02,  2.20s/it]\n",
      "Loss: 0.01255:  67%|██████▋   | 167/250 [06:10<03:02,  2.20s/it]\n",
      "Loss: 0.01311:  67%|██████▋   | 167/250 [06:10<03:02,  2.20s/it]\n",
      "Loss: 0.01083:  67%|██████▋   | 167/250 [06:10<03:02,  2.20s/it]\n",
      "Loss: 0.01150:  67%|██████▋   | 167/250 [06:10<03:02,  2.20s/it]\n",
      "Loss: 0.01261:  67%|██████▋   | 167/250 [06:10<03:02,  2.20s/it]\n",
      "Loss: 0.01034:  67%|██████▋   | 167/250 [06:10<03:02,  2.20s/it]\n",
      "Loss: 0.01067:  67%|██████▋   | 167/250 [06:11<03:02,  2.20s/it]\n",
      "Loss: 0.01116:  67%|██████▋   | 167/250 [06:11<03:02,  2.20s/it]\n",
      "Loss: 0.01133:  67%|██████▋   | 167/250 [06:11<03:02,  2.20s/it]\n",
      "Loss: 0.01219:  67%|██████▋   | 167/250 [06:11<03:02,  2.20s/it]\n",
      "Loss: 0.01086:  67%|██████▋   | 167/250 [06:11<03:02,  2.20s/it]\n",
      "Loss: 0.01132:  67%|██████▋   | 167/250 [06:11<03:02,  2.20s/it]\n",
      "Loss: 0.01132:  67%|██████▋   | 168/250 [06:11<03:01,  2.21s/it]\n",
      "Loss: 0.01242:  67%|██████▋   | 168/250 [06:11<03:01,  2.21s/it]\n",
      "Loss: 0.01109:  67%|██████▋   | 168/250 [06:11<03:01,  2.21s/it]\n",
      "Loss: 0.01114:  67%|██████▋   | 168/250 [06:11<03:01,  2.21s/it]\n",
      "Loss: 0.01238:  67%|██████▋   | 168/250 [06:12<03:01,  2.21s/it]\n",
      "Loss: 0.00898:  67%|██████▋   | 168/250 [06:12<03:01,  2.21s/it]\n",
      "Loss: 0.01007:  67%|██████▋   | 168/250 [06:12<03:01,  2.21s/it]\n",
      "Loss: 0.00978:  67%|██████▋   | 168/250 [06:12<03:01,  2.21s/it]\n",
      "Loss: 0.01176:  67%|██████▋   | 168/250 [06:12<03:01,  2.21s/it]\n",
      "Loss: 0.01070:  67%|██████▋   | 168/250 [06:12<03:01,  2.21s/it]\n",
      "Loss: 0.01218:  67%|██████▋   | 168/250 [06:12<03:01,  2.21s/it]\n",
      "Loss: 0.01148:  67%|██████▋   | 168/250 [06:12<03:01,  2.21s/it]\n",
      "Loss: 0.01055:  67%|██████▋   | 168/250 [06:12<03:01,  2.21s/it]\n",
      "Loss: 0.01060:  67%|██████▋   | 168/250 [06:12<03:01,  2.21s/it]\n",
      "Loss: 0.01048:  67%|██████▋   | 168/250 [06:13<03:01,  2.21s/it]\n",
      "Loss: 0.01148:  67%|██████▋   | 168/250 [06:13<03:01,  2.21s/it]\n",
      "Loss: 0.00969:  67%|██████▋   | 168/250 [06:13<03:01,  2.21s/it]\n",
      "Loss: 0.01168:  67%|██████▋   | 168/250 [06:13<03:01,  2.21s/it]\n",
      "Loss: 0.01116:  67%|██████▋   | 168/250 [06:13<03:01,  2.21s/it]\n",
      "Loss: 0.01009:  67%|██████▋   | 168/250 [06:13<03:01,  2.21s/it]\n",
      "Loss: 0.01090:  67%|██████▋   | 168/250 [06:13<03:01,  2.21s/it]\n",
      "Loss: 0.01141:  67%|██████▋   | 168/250 [06:13<03:01,  2.21s/it]\n",
      "Loss: 0.01141:  68%|██████▊   | 169/250 [06:13<02:58,  2.20s/it]\n",
      "Loss: 0.01099:  68%|██████▊   | 169/250 [06:13<02:58,  2.20s/it]\n",
      "Loss: 0.01116:  68%|██████▊   | 169/250 [06:13<02:58,  2.20s/it]\n",
      "Loss: 0.01283:  68%|██████▊   | 169/250 [06:14<02:58,  2.20s/it]\n",
      "Loss: 0.00996:  68%|██████▊   | 169/250 [06:14<02:58,  2.20s/it]\n",
      "Loss: 0.00969:  68%|██████▊   | 169/250 [06:14<02:58,  2.20s/it]\n",
      "Loss: 0.01170:  68%|██████▊   | 169/250 [06:14<02:58,  2.20s/it]\n",
      "Loss: 0.01208:  68%|██████▊   | 169/250 [06:14<02:58,  2.20s/it]\n",
      "Loss: 0.01084:  68%|██████▊   | 169/250 [06:14<02:58,  2.20s/it]\n",
      "Loss: 0.01020:  68%|██████▊   | 169/250 [06:14<02:58,  2.20s/it]\n",
      "Loss: 0.01230:  68%|██████▊   | 169/250 [06:14<02:58,  2.20s/it]\n",
      "Loss: 0.01136:  68%|██████▊   | 169/250 [06:14<02:58,  2.20s/it]\n",
      "Loss: 0.01234:  68%|██████▊   | 169/250 [06:15<02:58,  2.20s/it]\n",
      "Loss: 0.01216:  68%|██████▊   | 169/250 [06:15<02:58,  2.20s/it]\n",
      "Loss: 0.01173:  68%|██████▊   | 169/250 [06:15<02:58,  2.20s/it]\n",
      "Loss: 0.01168:  68%|██████▊   | 169/250 [06:15<02:58,  2.20s/it]\n",
      "Loss: 0.01134:  68%|██████▊   | 169/250 [06:15<02:58,  2.20s/it]\n",
      "Loss: 0.01195:  68%|██████▊   | 169/250 [06:15<02:58,  2.20s/it]\n",
      "Loss: 0.01047:  68%|██████▊   | 169/250 [06:15<02:58,  2.20s/it]\n",
      "Loss: 0.01161:  68%|██████▊   | 169/250 [06:15<02:58,  2.20s/it]\n",
      "Loss: 0.01115:  68%|██████▊   | 169/250 [06:15<02:58,  2.20s/it]\n",
      "Loss: 0.01157:  68%|██████▊   | 169/250 [06:15<02:58,  2.20s/it]\n",
      "Loss: 0.01157:  68%|██████▊   | 170/250 [06:15<02:56,  2.20s/it]\n",
      "Loss: 0.01060:  68%|██████▊   | 170/250 [06:16<02:56,  2.20s/it]\n",
      "Loss: 0.01111:  68%|██████▊   | 170/250 [06:16<02:56,  2.20s/it]\n",
      "Loss: 0.01307:  68%|██████▊   | 170/250 [06:16<02:56,  2.20s/it]\n",
      "Loss: 0.00961:  68%|██████▊   | 170/250 [06:16<02:56,  2.20s/it]\n",
      "Loss: 0.00947:  68%|██████▊   | 170/250 [06:16<02:56,  2.20s/it]\n",
      "Loss: 0.01156:  68%|██████▊   | 170/250 [06:16<02:56,  2.20s/it]\n",
      "Loss: 0.00938:  68%|██████▊   | 170/250 [06:16<02:56,  2.20s/it]\n",
      "Loss: 0.01439:  68%|██████▊   | 170/250 [06:16<02:56,  2.20s/it]\n",
      "Loss: 0.00966:  68%|██████▊   | 170/250 [06:16<02:56,  2.20s/it]\n",
      "Loss: 0.01260:  68%|██████▊   | 170/250 [06:17<02:56,  2.20s/it]\n",
      "Loss: 0.01293:  68%|██████▊   | 170/250 [06:17<02:56,  2.20s/it]\n",
      "Loss: 0.01095:  68%|██████▊   | 170/250 [06:17<02:56,  2.20s/it]\n",
      "Loss: 0.01161:  68%|██████▊   | 170/250 [06:17<02:56,  2.20s/it]\n",
      "Loss: 0.01282:  68%|██████▊   | 170/250 [06:17<02:56,  2.20s/it]\n",
      "Loss: 0.01110:  68%|██████▊   | 170/250 [06:17<02:56,  2.20s/it]\n",
      "Loss: 0.01046:  68%|██████▊   | 170/250 [06:17<02:56,  2.20s/it]\n",
      "Loss: 0.01098:  68%|██████▊   | 170/250 [06:17<02:56,  2.20s/it]\n",
      "Loss: 0.01129:  68%|██████▊   | 170/250 [06:17<02:56,  2.20s/it]\n",
      "Loss: 0.01357:  68%|██████▊   | 170/250 [06:17<02:56,  2.20s/it]\n",
      "Loss: 0.01130:  68%|██████▊   | 170/250 [06:18<02:56,  2.20s/it]\n",
      "Loss: 0.01124:  68%|██████▊   | 170/250 [06:18<02:56,  2.20s/it]\n",
      "Loss: 0.01124:  68%|██████▊   | 171/250 [06:18<02:54,  2.21s/it]\n",
      "Loss: 0.01068:  68%|██████▊   | 171/250 [06:18<02:54,  2.21s/it]\n",
      "Loss: 0.01130:  68%|██████▊   | 171/250 [06:18<02:54,  2.21s/it]\n",
      "Loss: 0.01106:  68%|██████▊   | 171/250 [06:18<02:54,  2.21s/it]\n",
      "Loss: 0.01226:  68%|██████▊   | 171/250 [06:18<02:54,  2.21s/it]\n",
      "Loss: 0.01228:  68%|██████▊   | 171/250 [06:18<02:54,  2.21s/it]\n",
      "Loss: 0.01138:  68%|██████▊   | 171/250 [06:18<02:54,  2.21s/it]\n",
      "Loss: 0.01206:  68%|██████▊   | 171/250 [06:18<02:54,  2.21s/it]\n",
      "Loss: 0.01075:  68%|██████▊   | 171/250 [06:19<02:54,  2.21s/it]\n",
      "Loss: 0.01284:  68%|██████▊   | 171/250 [06:19<02:54,  2.21s/it]\n",
      "Loss: 0.01170:  68%|██████▊   | 171/250 [06:19<02:54,  2.21s/it]\n",
      "Loss: 0.01042:  68%|██████▊   | 171/250 [06:19<02:54,  2.21s/it]\n",
      "Loss: 0.01077:  68%|██████▊   | 171/250 [06:19<02:54,  2.21s/it]\n",
      "Loss: 0.01125:  68%|██████▊   | 171/250 [06:19<02:54,  2.21s/it]\n",
      "Loss: 0.01008:  68%|██████▊   | 171/250 [06:19<02:54,  2.21s/it]\n",
      "Loss: 0.01063:  68%|██████▊   | 171/250 [06:19<02:54,  2.21s/it]\n",
      "Loss: 0.01215:  68%|██████▊   | 171/250 [06:19<02:54,  2.21s/it]\n",
      "Loss: 0.01023:  68%|██████▊   | 171/250 [06:20<02:54,  2.21s/it]\n",
      "Loss: 0.01248:  68%|██████▊   | 171/250 [06:20<02:54,  2.21s/it]\n",
      "Loss: 0.01206:  68%|██████▊   | 171/250 [06:20<02:54,  2.21s/it]\n",
      "Loss: 0.01090:  68%|██████▊   | 171/250 [06:20<02:54,  2.21s/it]\n",
      "Loss: 0.01197:  68%|██████▊   | 171/250 [06:20<02:54,  2.21s/it]\n",
      "Loss: 0.01197:  69%|██████▉   | 172/250 [06:20<02:52,  2.21s/it]\n",
      "Loss: 0.01109:  69%|██████▉   | 172/250 [06:20<02:52,  2.21s/it]\n",
      "Loss: 0.01087:  69%|██████▉   | 172/250 [06:20<02:52,  2.21s/it]\n",
      "Loss: 0.01127:  69%|██████▉   | 172/250 [06:20<02:52,  2.21s/it]\n",
      "Loss: 0.01230:  69%|██████▉   | 172/250 [06:20<02:52,  2.21s/it]\n",
      "Loss: 0.01242:  69%|██████▉   | 172/250 [06:20<02:52,  2.21s/it]\n",
      "Loss: 0.01156:  69%|██████▉   | 172/250 [06:21<02:52,  2.21s/it]\n",
      "Loss: 0.01195:  69%|██████▉   | 172/250 [06:21<02:52,  2.21s/it]\n",
      "Loss: 0.01092:  69%|██████▉   | 172/250 [06:21<02:52,  2.21s/it]\n",
      "Loss: 0.01067:  69%|██████▉   | 172/250 [06:21<02:52,  2.21s/it]\n",
      "Loss: 0.01234:  69%|██████▉   | 172/250 [06:21<02:52,  2.21s/it]\n",
      "Loss: 0.01149:  69%|██████▉   | 172/250 [06:21<02:52,  2.21s/it]\n",
      "Loss: 0.01176:  69%|██████▉   | 172/250 [06:21<02:52,  2.21s/it]\n",
      "Loss: 0.01231:  69%|██████▉   | 172/250 [06:21<02:52,  2.21s/it]\n",
      "Loss: 0.01166:  69%|██████▉   | 172/250 [06:21<02:52,  2.21s/it]\n",
      "Loss: 0.01148:  69%|██████▉   | 172/250 [06:22<02:52,  2.21s/it]\n",
      "Loss: 0.01133:  69%|██████▉   | 172/250 [06:22<02:52,  2.21s/it]\n",
      "Loss: 0.01189:  69%|██████▉   | 172/250 [06:22<02:52,  2.21s/it]\n",
      "Loss: 0.01043:  69%|██████▉   | 172/250 [06:22<02:52,  2.21s/it]\n",
      "Loss: 0.01109:  69%|██████▉   | 172/250 [06:22<02:52,  2.21s/it]\n",
      "Loss: 0.01117:  69%|██████▉   | 172/250 [06:22<02:52,  2.21s/it]\n",
      "Loss: 0.01128:  69%|██████▉   | 172/250 [06:22<02:52,  2.21s/it]\n",
      "Loss: 0.01128:  69%|██████▉   | 173/250 [06:22<02:50,  2.22s/it]\n",
      "Loss: 0.01106:  69%|██████▉   | 173/250 [06:22<02:50,  2.22s/it]\n",
      "Loss: 0.01108:  69%|██████▉   | 173/250 [06:22<02:50,  2.22s/it]\n",
      "Loss: 0.01276:  69%|██████▉   | 173/250 [06:22<02:50,  2.22s/it]\n",
      "Loss: 0.00983:  69%|██████▉   | 173/250 [06:23<02:50,  2.22s/it]\n",
      "Loss: 0.00938:  69%|██████▉   | 173/250 [06:23<02:50,  2.22s/it]\n",
      "Loss: 0.01169:  69%|██████▉   | 173/250 [06:23<02:50,  2.22s/it]\n",
      "Loss: 0.00920:  69%|██████▉   | 173/250 [06:23<02:50,  2.22s/it]\n",
      "Loss: 0.01429:  69%|██████▉   | 173/250 [06:23<02:50,  2.22s/it]\n",
      "Loss: 0.00960:  69%|██████▉   | 173/250 [06:23<02:50,  2.22s/it]\n",
      "Loss: 0.01278:  69%|██████▉   | 173/250 [06:23<02:50,  2.22s/it]\n",
      "Loss: 0.01260:  69%|██████▉   | 173/250 [06:23<02:50,  2.22s/it]\n",
      "Loss: 0.01074:  69%|██████▉   | 173/250 [06:23<02:50,  2.22s/it]\n",
      "Loss: 0.01166:  69%|██████▉   | 173/250 [06:24<02:50,  2.22s/it]\n",
      "Loss: 0.01287:  69%|██████▉   | 173/250 [06:24<02:50,  2.22s/it]\n",
      "Loss: 0.01052:  69%|██████▉   | 173/250 [06:24<02:50,  2.22s/it]\n",
      "Loss: 0.01003:  69%|██████▉   | 173/250 [06:24<02:50,  2.22s/it]\n",
      "Loss: 0.01112:  69%|██████▉   | 173/250 [06:24<02:50,  2.22s/it]\n",
      "Loss: 0.01123:  69%|██████▉   | 173/250 [06:24<02:50,  2.22s/it]\n",
      "Loss: 0.01346:  69%|██████▉   | 173/250 [06:24<02:50,  2.22s/it]\n",
      "Loss: 0.01060:  69%|██████▉   | 173/250 [06:24<02:50,  2.22s/it]\n",
      "Loss: 0.01127:  69%|██████▉   | 173/250 [06:24<02:50,  2.22s/it]\n",
      "Loss: 0.01127:  70%|██████▉   | 174/250 [06:24<02:48,  2.22s/it]\n",
      "Loss: 0.01063:  70%|██████▉   | 174/250 [06:24<02:48,  2.22s/it]\n",
      "Loss: 0.01111:  70%|██████▉   | 174/250 [06:25<02:48,  2.22s/it]\n",
      "Loss: 0.01118:  70%|██████▉   | 174/250 [06:25<02:48,  2.22s/it]\n",
      "Loss: 0.01227:  70%|██████▉   | 174/250 [06:25<02:48,  2.22s/it]\n",
      "Loss: 0.01232:  70%|██████▉   | 174/250 [06:25<02:48,  2.22s/it]\n",
      "Loss: 0.01122:  70%|██████▉   | 174/250 [06:25<02:48,  2.22s/it]\n",
      "Loss: 0.01209:  70%|██████▉   | 174/250 [06:25<02:48,  2.22s/it]\n",
      "Loss: 0.01080:  70%|██████▉   | 174/250 [06:25<02:48,  2.22s/it]\n",
      "Loss: 0.01283:  70%|██████▉   | 174/250 [06:25<02:48,  2.22s/it]\n",
      "Loss: 0.01160:  70%|██████▉   | 174/250 [06:25<02:48,  2.22s/it]\n",
      "Loss: 0.01037:  70%|██████▉   | 174/250 [06:26<02:48,  2.22s/it]\n",
      "Loss: 0.01041:  70%|██████▉   | 174/250 [06:26<02:48,  2.22s/it]\n",
      "Loss: 0.01088:  70%|██████▉   | 174/250 [06:26<02:48,  2.22s/it]\n",
      "Loss: 0.00974:  70%|██████▉   | 174/250 [06:26<02:48,  2.22s/it]\n",
      "Loss: 0.01038:  70%|██████▉   | 174/250 [06:26<02:48,  2.22s/it]\n",
      "Loss: 0.01214:  70%|██████▉   | 174/250 [06:26<02:48,  2.22s/it]\n",
      "Loss: 0.01024:  70%|██████▉   | 174/250 [06:26<02:48,  2.22s/it]\n",
      "Loss: 0.01244:  70%|██████▉   | 174/250 [06:26<02:48,  2.22s/it]\n",
      "Loss: 0.01203:  70%|██████▉   | 174/250 [06:26<02:48,  2.22s/it]\n",
      "Loss: 0.01108:  70%|██████▉   | 174/250 [06:27<02:48,  2.22s/it]\n",
      "Loss: 0.01142:  70%|██████▉   | 174/250 [06:27<02:48,  2.22s/it]\n",
      "Loss: 0.01142:  70%|███████   | 175/250 [06:27<02:47,  2.23s/it]\n",
      "Loss: 0.01091:  70%|███████   | 175/250 [06:27<02:47,  2.23s/it]\n",
      "Loss: 0.01123:  70%|███████   | 175/250 [06:27<02:47,  2.23s/it]\n",
      "Loss: 0.01264:  70%|███████   | 175/250 [06:27<02:47,  2.23s/it]\n",
      "Loss: 0.00939:  70%|███████   | 175/250 [06:27<02:47,  2.23s/it]\n",
      "Loss: 0.00950:  70%|███████   | 175/250 [06:27<02:47,  2.23s/it]\n",
      "Loss: 0.01044:  70%|███████   | 175/250 [06:27<02:47,  2.23s/it]\n",
      "Loss: 0.01138:  70%|███████   | 175/250 [06:27<02:47,  2.23s/it]\n",
      "Loss: 0.01397:  70%|███████   | 175/250 [06:27<02:47,  2.23s/it]\n",
      "Loss: 0.00958:  70%|███████   | 175/250 [06:28<02:47,  2.23s/it]\n",
      "Loss: 0.01225:  70%|███████   | 175/250 [06:28<02:47,  2.23s/it]\n",
      "Loss: 0.01149:  70%|███████   | 175/250 [06:28<02:47,  2.23s/it]\n",
      "Loss: 0.01048:  70%|███████   | 175/250 [06:28<02:47,  2.23s/it]\n",
      "Loss: 0.01076:  70%|███████   | 175/250 [06:28<02:47,  2.23s/it]\n",
      "Loss: 0.00943:  70%|███████   | 175/250 [06:28<02:47,  2.23s/it]\n",
      "Loss: 0.00995:  70%|███████   | 175/250 [06:28<02:47,  2.23s/it]\n",
      "Loss: 0.01023:  70%|███████   | 175/250 [06:28<02:47,  2.23s/it]\n",
      "Loss: 0.01136:  70%|███████   | 175/250 [06:28<02:47,  2.23s/it]\n",
      "Loss: 0.01079:  70%|███████   | 175/250 [06:29<02:47,  2.23s/it]\n",
      "Loss: 0.01217:  70%|███████   | 175/250 [06:29<02:47,  2.23s/it]\n",
      "Loss: 0.01082:  70%|███████   | 175/250 [06:29<02:47,  2.23s/it]\n",
      "Loss: 0.01134:  70%|███████   | 175/250 [06:29<02:47,  2.23s/it]\n",
      "Loss: 0.01134:  70%|███████   | 176/250 [06:29<02:44,  2.22s/it]\n",
      "Loss: 0.01140:  70%|███████   | 176/250 [06:29<02:44,  2.22s/it]\n",
      "Loss: 0.01065:  70%|███████   | 176/250 [06:29<02:44,  2.22s/it]\n",
      "Loss: 0.01089:  70%|███████   | 176/250 [06:29<02:44,  2.22s/it]\n",
      "Loss: 0.01219:  70%|███████   | 176/250 [06:29<02:44,  2.22s/it]\n",
      "Loss: 0.01208:  70%|███████   | 176/250 [06:29<02:44,  2.22s/it]\n",
      "Loss: 0.01107:  70%|███████   | 176/250 [06:29<02:44,  2.22s/it]\n",
      "Loss: 0.01208:  70%|███████   | 176/250 [06:30<02:44,  2.22s/it]\n",
      "Loss: 0.01087:  70%|███████   | 176/250 [06:30<02:44,  2.22s/it]\n",
      "Loss: 0.01267:  70%|███████   | 176/250 [06:30<02:44,  2.22s/it]\n",
      "Loss: 0.01158:  70%|███████   | 176/250 [06:30<02:44,  2.22s/it]\n",
      "Loss: 0.01063:  70%|███████   | 176/250 [06:30<02:44,  2.22s/it]\n",
      "Loss: 0.01170:  70%|███████   | 176/250 [06:30<02:44,  2.22s/it]\n",
      "Loss: 0.01212:  70%|███████   | 176/250 [06:30<02:44,  2.22s/it]\n",
      "Loss: 0.01122:  70%|███████   | 176/250 [06:30<02:44,  2.22s/it]\n",
      "Loss: 0.01127:  70%|███████   | 176/250 [06:30<02:44,  2.22s/it]\n",
      "Loss: 0.01070:  70%|███████   | 176/250 [06:31<02:44,  2.22s/it]\n",
      "Loss: 0.01177:  70%|███████   | 176/250 [06:31<02:44,  2.22s/it]\n",
      "Loss: 0.01061:  70%|███████   | 176/250 [06:31<02:44,  2.22s/it]\n",
      "Loss: 0.01105:  70%|███████   | 176/250 [06:31<02:44,  2.22s/it]\n",
      "Loss: 0.01093:  70%|███████   | 176/250 [06:31<02:44,  2.22s/it]\n",
      "Loss: 0.01159:  70%|███████   | 176/250 [06:31<02:44,  2.22s/it]\n",
      "Loss: 0.01159:  71%|███████   | 177/250 [06:31<02:41,  2.21s/it]\n",
      "Loss: 0.01052:  71%|███████   | 177/250 [06:31<02:41,  2.21s/it]\n",
      "Loss: 0.01100:  71%|███████   | 177/250 [06:31<02:41,  2.21s/it]\n",
      "Loss: 0.01290:  71%|███████   | 177/250 [06:31<02:41,  2.21s/it]\n",
      "Loss: 0.00936:  71%|███████   | 177/250 [06:31<02:41,  2.21s/it]\n",
      "Loss: 0.00945:  71%|███████   | 177/250 [06:32<02:41,  2.21s/it]\n",
      "Loss: 0.01141:  71%|███████   | 177/250 [06:32<02:41,  2.21s/it]\n",
      "Loss: 0.00915:  71%|███████   | 177/250 [06:32<02:41,  2.21s/it]\n",
      "Loss: 0.01390:  71%|███████   | 177/250 [06:32<02:41,  2.21s/it]\n",
      "Loss: 0.01001:  71%|███████   | 177/250 [06:32<02:41,  2.21s/it]\n",
      "Loss: 0.01237:  71%|███████   | 177/250 [06:32<02:41,  2.21s/it]\n",
      "Loss: 0.01253:  71%|███████   | 177/250 [06:32<02:41,  2.21s/it]\n",
      "Loss: 0.01092:  71%|███████   | 177/250 [06:32<02:41,  2.21s/it]\n",
      "Loss: 0.01192:  71%|███████   | 177/250 [06:32<02:41,  2.21s/it]\n",
      "Loss: 0.01252:  71%|███████   | 177/250 [06:32<02:41,  2.21s/it]\n",
      "Loss: 0.01055:  71%|███████   | 177/250 [06:33<02:41,  2.21s/it]\n",
      "Loss: 0.00984:  71%|███████   | 177/250 [06:33<02:41,  2.21s/it]\n",
      "Loss: 0.01101:  71%|███████   | 177/250 [06:33<02:41,  2.21s/it]\n",
      "Loss: 0.01122:  71%|███████   | 177/250 [06:33<02:41,  2.21s/it]\n",
      "Loss: 0.01414:  71%|███████   | 177/250 [06:33<02:41,  2.21s/it]\n",
      "Loss: 0.01113:  71%|███████   | 177/250 [06:33<02:41,  2.21s/it]\n",
      "Loss: 0.01128:  71%|███████   | 177/250 [06:33<02:41,  2.21s/it]\n",
      "Loss: 0.01128:  71%|███████   | 178/250 [06:33<02:39,  2.21s/it]\n",
      "Loss: 0.01043:  71%|███████   | 178/250 [06:33<02:39,  2.21s/it]\n",
      "Loss: 0.01089:  71%|███████   | 178/250 [06:33<02:39,  2.21s/it]\n",
      "Loss: 0.01267:  71%|███████   | 178/250 [06:34<02:39,  2.21s/it]\n",
      "Loss: 0.00959:  71%|███████   | 178/250 [06:34<02:39,  2.21s/it]\n",
      "Loss: 0.00942:  71%|███████   | 178/250 [06:34<02:39,  2.21s/it]\n",
      "Loss: 0.01147:  71%|███████   | 178/250 [06:34<02:39,  2.21s/it]\n",
      "Loss: 0.01195:  71%|███████   | 178/250 [06:34<02:39,  2.21s/it]\n",
      "Loss: 0.01056:  71%|███████   | 178/250 [06:34<02:39,  2.21s/it]\n",
      "Loss: 0.01313:  71%|███████   | 178/250 [06:34<02:39,  2.21s/it]\n",
      "Loss: 0.01154:  71%|███████   | 178/250 [06:34<02:39,  2.21s/it]\n",
      "Loss: 0.01036:  71%|███████   | 178/250 [06:34<02:39,  2.21s/it]\n",
      "Loss: 0.01055:  71%|███████   | 178/250 [06:34<02:39,  2.21s/it]\n",
      "Loss: 0.01099:  71%|███████   | 178/250 [06:35<02:39,  2.21s/it]\n",
      "Loss: 0.00969:  71%|███████   | 178/250 [06:35<02:39,  2.21s/it]\n",
      "Loss: 0.01029:  71%|███████   | 178/250 [06:35<02:39,  2.21s/it]\n",
      "Loss: 0.01212:  71%|███████   | 178/250 [06:35<02:39,  2.21s/it]\n",
      "Loss: 0.01195:  71%|███████   | 178/250 [06:35<02:39,  2.21s/it]\n",
      "Loss: 0.01037:  71%|███████   | 178/250 [06:35<02:39,  2.21s/it]\n",
      "Loss: 0.01152:  71%|███████   | 178/250 [06:35<02:39,  2.21s/it]\n",
      "Loss: 0.01105:  71%|███████   | 178/250 [06:35<02:39,  2.21s/it]\n",
      "Loss: 0.01116:  71%|███████   | 178/250 [06:35<02:39,  2.21s/it]\n",
      "Loss: 0.01116:  72%|███████▏  | 179/250 [06:35<02:36,  2.21s/it]\n",
      "Loss: 0.01076:  72%|███████▏  | 179/250 [06:36<02:36,  2.21s/it]\n",
      "Loss: 0.01104:  72%|███████▏  | 179/250 [06:36<02:36,  2.21s/it]\n",
      "Loss: 0.01252:  72%|███████▏  | 179/250 [06:36<02:36,  2.21s/it]\n",
      "Loss: 0.00923:  72%|███████▏  | 179/250 [06:36<02:36,  2.21s/it]\n",
      "Loss: 0.00962:  72%|███████▏  | 179/250 [06:36<02:36,  2.21s/it]\n",
      "Loss: 0.01143:  72%|███████▏  | 179/250 [06:36<02:36,  2.21s/it]\n",
      "Loss: 0.00947:  72%|███████▏  | 179/250 [06:36<02:36,  2.21s/it]\n",
      "Loss: 0.01424:  72%|███████▏  | 179/250 [06:36<02:36,  2.21s/it]\n",
      "Loss: 0.00976:  72%|███████▏  | 179/250 [06:36<02:36,  2.21s/it]\n",
      "Loss: 0.01127:  72%|███████▏  | 179/250 [06:36<02:36,  2.21s/it]\n",
      "Loss: 0.01028:  72%|███████▏  | 179/250 [06:37<02:36,  2.21s/it]\n",
      "Loss: 0.01048:  72%|███████▏  | 179/250 [06:37<02:36,  2.21s/it]\n",
      "Loss: 0.01097:  72%|███████▏  | 179/250 [06:37<02:36,  2.21s/it]\n",
      "Loss: 0.00993:  72%|███████▏  | 179/250 [06:37<02:36,  2.21s/it]\n",
      "Loss: 0.01007:  72%|███████▏  | 179/250 [06:37<02:36,  2.21s/it]\n",
      "Loss: 0.01210:  72%|███████▏  | 179/250 [06:37<02:36,  2.21s/it]\n",
      "Loss: 0.00995:  72%|███████▏  | 179/250 [06:37<02:36,  2.21s/it]\n",
      "Loss: 0.01245:  72%|███████▏  | 179/250 [06:37<02:36,  2.21s/it]\n",
      "Loss: 0.01200:  72%|███████▏  | 179/250 [06:37<02:36,  2.21s/it]\n",
      "Loss: 0.01062:  72%|███████▏  | 179/250 [06:38<02:36,  2.21s/it]\n",
      "Loss: 0.01124:  72%|███████▏  | 179/250 [06:38<02:36,  2.21s/it]\n",
      "Loss: 0.01124:  72%|███████▏  | 180/250 [06:38<02:34,  2.21s/it]\n",
      "Loss: 0.01141:  72%|███████▏  | 180/250 [06:38<02:34,  2.21s/it]\n",
      "Loss: 0.01086:  72%|███████▏  | 180/250 [06:38<02:34,  2.21s/it]\n",
      "Loss: 0.01075:  72%|███████▏  | 180/250 [06:38<02:34,  2.21s/it]\n",
      "Loss: 0.01062:  72%|███████▏  | 180/250 [06:38<02:34,  2.21s/it]\n",
      "Loss: 0.00870:  72%|███████▏  | 180/250 [06:38<02:34,  2.21s/it]\n",
      "Loss: 0.00968:  72%|███████▏  | 180/250 [06:38<02:34,  2.21s/it]\n",
      "Loss: 0.00992:  72%|███████▏  | 180/250 [06:38<02:34,  2.21s/it]\n",
      "Loss: 0.01161:  72%|███████▏  | 180/250 [06:39<02:34,  2.21s/it]\n",
      "Loss: 0.01049:  72%|███████▏  | 180/250 [06:39<02:34,  2.21s/it]\n",
      "Loss: 0.01197:  72%|███████▏  | 180/250 [06:39<02:34,  2.21s/it]\n",
      "Loss: 0.01125:  72%|███████▏  | 180/250 [06:39<02:34,  2.21s/it]\n",
      "Loss: 0.01179:  72%|███████▏  | 180/250 [06:39<02:34,  2.21s/it]\n",
      "Loss: 0.01208:  72%|███████▏  | 180/250 [06:39<02:34,  2.21s/it]\n",
      "Loss: 0.01156:  72%|███████▏  | 180/250 [06:39<02:34,  2.21s/it]\n",
      "Loss: 0.01119:  72%|███████▏  | 180/250 [06:39<02:34,  2.21s/it]\n",
      "Loss: 0.01078:  72%|███████▏  | 180/250 [06:39<02:34,  2.21s/it]\n",
      "Loss: 0.01164:  72%|███████▏  | 180/250 [06:39<02:34,  2.21s/it]\n",
      "Loss: 0.01075:  72%|███████▏  | 180/250 [06:40<02:34,  2.21s/it]\n",
      "Loss: 0.01107:  72%|███████▏  | 180/250 [06:40<02:34,  2.21s/it]\n",
      "Loss: 0.01094:  72%|███████▏  | 180/250 [06:40<02:34,  2.21s/it]\n",
      "Loss: 0.01132:  72%|███████▏  | 180/250 [06:40<02:34,  2.21s/it]\n",
      "Loss: 0.01132:  72%|███████▏  | 181/250 [06:40<02:31,  2.20s/it]\n",
      "Loss: 0.01026:  72%|███████▏  | 181/250 [06:40<02:31,  2.20s/it]\n",
      "Loss: 0.01097:  72%|███████▏  | 181/250 [06:40<02:31,  2.20s/it]\n",
      "Loss: 0.01244:  72%|███████▏  | 181/250 [06:40<02:31,  2.20s/it]\n",
      "Loss: 0.00940:  72%|███████▏  | 181/250 [06:40<02:31,  2.20s/it]\n",
      "Loss: 0.00938:  72%|███████▏  | 181/250 [06:40<02:31,  2.20s/it]\n",
      "Loss: 0.01173:  72%|███████▏  | 181/250 [06:40<02:31,  2.20s/it]\n",
      "Loss: 0.00885:  72%|███████▏  | 181/250 [06:41<02:31,  2.20s/it]\n",
      "Loss: 0.01381:  72%|███████▏  | 181/250 [06:41<02:31,  2.20s/it]\n",
      "Loss: 0.00970:  72%|███████▏  | 181/250 [06:41<02:31,  2.20s/it]\n",
      "Loss: 0.01210:  72%|███████▏  | 181/250 [06:41<02:31,  2.20s/it]\n",
      "Loss: 0.01269:  72%|███████▏  | 181/250 [06:41<02:31,  2.20s/it]\n",
      "Loss: 0.01056:  72%|███████▏  | 181/250 [06:41<02:31,  2.20s/it]\n",
      "Loss: 0.01149:  72%|███████▏  | 181/250 [06:41<02:31,  2.20s/it]\n",
      "Loss: 0.01249:  72%|███████▏  | 181/250 [06:41<02:31,  2.20s/it]\n",
      "Loss: 0.01131:  72%|███████▏  | 181/250 [06:41<02:31,  2.20s/it]\n",
      "Loss: 0.01006:  72%|███████▏  | 181/250 [06:41<02:31,  2.20s/it]\n",
      "Loss: 0.01021:  72%|███████▏  | 181/250 [06:42<02:31,  2.20s/it]\n",
      "Loss: 0.01231:  72%|███████▏  | 181/250 [06:42<02:31,  2.20s/it]\n",
      "Loss: 0.01200:  72%|███████▏  | 181/250 [06:42<02:31,  2.20s/it]\n",
      "Loss: 0.01049:  72%|███████▏  | 181/250 [06:42<02:31,  2.20s/it]\n",
      "Loss: 0.01133:  72%|███████▏  | 181/250 [06:42<02:31,  2.20s/it]\n",
      "Loss: 0.01133:  73%|███████▎  | 182/250 [06:42<02:29,  2.19s/it]\n",
      "Loss: 0.01027:  73%|███████▎  | 182/250 [06:42<02:29,  2.19s/it]\n",
      "Loss: 0.01132:  73%|███████▎  | 182/250 [06:42<02:29,  2.19s/it]\n",
      "Loss: 0.01255:  73%|███████▎  | 182/250 [06:42<02:29,  2.19s/it]\n",
      "Loss: 0.00952:  73%|███████▎  | 182/250 [06:42<02:29,  2.19s/it]\n",
      "Loss: 0.00949:  73%|███████▎  | 182/250 [06:43<02:29,  2.19s/it]\n",
      "Loss: 0.01124:  73%|███████▎  | 182/250 [06:43<02:29,  2.19s/it]\n",
      "Loss: 0.00920:  73%|███████▎  | 182/250 [06:43<02:29,  2.19s/it]\n",
      "Loss: 0.01385:  73%|███████▎  | 182/250 [06:43<02:29,  2.19s/it]\n",
      "Loss: 0.00976:  73%|███████▎  | 182/250 [06:43<02:29,  2.19s/it]\n",
      "Loss: 0.01270:  73%|███████▎  | 182/250 [06:43<02:29,  2.19s/it]\n",
      "Loss: 0.01254:  73%|███████▎  | 182/250 [06:43<02:29,  2.19s/it]\n",
      "Loss: 0.01052:  73%|███████▎  | 182/250 [06:43<02:29,  2.19s/it]\n",
      "Loss: 0.01184:  73%|███████▎  | 182/250 [06:43<02:29,  2.19s/it]\n",
      "Loss: 0.01261:  73%|███████▎  | 182/250 [06:43<02:29,  2.19s/it]\n",
      "Loss: 0.01036:  73%|███████▎  | 182/250 [06:44<02:29,  2.19s/it]\n",
      "Loss: 0.00987:  73%|███████▎  | 182/250 [06:44<02:29,  2.19s/it]\n",
      "Loss: 0.01115:  73%|███████▎  | 182/250 [06:44<02:29,  2.19s/it]\n",
      "Loss: 0.01243:  73%|███████▎  | 182/250 [06:44<02:29,  2.19s/it]\n",
      "Loss: 0.01219:  73%|███████▎  | 182/250 [06:44<02:29,  2.19s/it]\n",
      "Loss: 0.01065:  73%|███████▎  | 182/250 [06:44<02:29,  2.19s/it]\n",
      "Loss: 0.01107:  73%|███████▎  | 182/250 [06:44<02:29,  2.19s/it]\n",
      "Loss: 0.01107:  73%|███████▎  | 183/250 [06:44<02:27,  2.19s/it]\n",
      "Loss: 0.01031:  73%|███████▎  | 183/250 [06:44<02:27,  2.19s/it]\n",
      "Loss: 0.01112:  73%|███████▎  | 183/250 [06:44<02:27,  2.19s/it]\n",
      "Loss: 0.01255:  73%|███████▎  | 183/250 [06:44<02:27,  2.19s/it]\n",
      "Loss: 0.00962:  73%|███████▎  | 183/250 [06:45<02:27,  2.19s/it]\n",
      "Loss: 0.00965:  73%|███████▎  | 183/250 [06:45<02:27,  2.19s/it]\n",
      "Loss: 0.01145:  73%|███████▎  | 183/250 [06:45<02:27,  2.19s/it]\n",
      "Loss: 0.00887:  73%|███████▎  | 183/250 [06:45<02:27,  2.19s/it]\n",
      "Loss: 0.01381:  73%|███████▎  | 183/250 [06:45<02:27,  2.19s/it]\n",
      "Loss: 0.00960:  73%|███████▎  | 183/250 [06:45<02:27,  2.19s/it]\n",
      "Loss: 0.01239:  73%|███████▎  | 183/250 [06:45<02:27,  2.19s/it]\n",
      "Loss: 0.01285:  73%|███████▎  | 183/250 [06:45<02:27,  2.19s/it]\n",
      "Loss: 0.01040:  73%|███████▎  | 183/250 [06:45<02:27,  2.19s/it]\n",
      "Loss: 0.01151:  73%|███████▎  | 183/250 [06:46<02:27,  2.19s/it]\n",
      "Loss: 0.01262:  73%|███████▎  | 183/250 [06:46<02:27,  2.19s/it]\n",
      "Loss: 0.01022:  73%|███████▎  | 183/250 [06:46<02:27,  2.19s/it]\n",
      "Loss: 0.00987:  73%|███████▎  | 183/250 [06:46<02:27,  2.19s/it]\n",
      "Loss: 0.01095:  73%|███████▎  | 183/250 [06:46<02:27,  2.19s/it]\n",
      "Loss: 0.01101:  73%|███████▎  | 183/250 [06:46<02:27,  2.19s/it]\n",
      "Loss: 0.01335:  73%|███████▎  | 183/250 [06:46<02:27,  2.19s/it]\n",
      "Loss: 0.01118:  73%|███████▎  | 183/250 [06:46<02:27,  2.19s/it]\n",
      "Loss: 0.01125:  73%|███████▎  | 183/250 [06:46<02:27,  2.19s/it]\n",
      "Loss: 0.01125:  74%|███████▎  | 184/250 [06:46<02:25,  2.21s/it]\n",
      "Loss: 0.01034:  74%|███████▎  | 184/250 [06:47<02:25,  2.21s/it]\n",
      "Loss: 0.01129:  74%|███████▎  | 184/250 [06:47<02:25,  2.21s/it]\n",
      "Loss: 0.01225:  74%|███████▎  | 184/250 [06:47<02:25,  2.21s/it]\n",
      "Loss: 0.00878:  74%|███████▎  | 184/250 [06:47<02:25,  2.21s/it]\n",
      "Loss: 0.00910:  74%|███████▎  | 184/250 [06:47<02:25,  2.21s/it]\n",
      "Loss: 0.01127:  74%|███████▎  | 184/250 [06:47<02:25,  2.21s/it]\n",
      "Loss: 0.00971:  74%|███████▎  | 184/250 [06:47<02:25,  2.21s/it]\n",
      "Loss: 0.00984:  74%|███████▎  | 184/250 [06:47<02:25,  2.21s/it]\n",
      "Loss: 0.01198:  74%|███████▎  | 184/250 [06:47<02:25,  2.21s/it]\n",
      "Loss: 0.01236:  74%|███████▎  | 184/250 [06:48<02:25,  2.21s/it]\n",
      "Loss: 0.01187:  74%|███████▎  | 184/250 [06:48<02:25,  2.21s/it]\n",
      "Loss: 0.01098:  74%|███████▎  | 184/250 [06:48<02:25,  2.21s/it]\n",
      "Loss: 0.01132:  74%|███████▎  | 184/250 [06:48<02:25,  2.21s/it]\n",
      "Loss: 0.01031:  74%|███████▎  | 184/250 [06:48<02:25,  2.21s/it]\n",
      "Loss: 0.00933:  74%|███████▎  | 184/250 [06:48<02:25,  2.21s/it]\n",
      "Loss: 0.00971:  74%|███████▎  | 184/250 [06:48<02:25,  2.21s/it]\n",
      "Loss: 0.01007:  74%|███████▎  | 184/250 [06:48<02:25,  2.21s/it]\n",
      "Loss: 0.01109:  74%|███████▎  | 184/250 [06:48<02:25,  2.21s/it]\n",
      "Loss: 0.01063:  74%|███████▎  | 184/250 [06:49<02:25,  2.21s/it]\n",
      "Loss: 0.01197:  74%|███████▎  | 184/250 [06:49<02:25,  2.21s/it]\n",
      "Loss: 0.01069:  74%|███████▎  | 184/250 [06:49<02:25,  2.21s/it]\n",
      "Loss: 0.01113:  74%|███████▎  | 184/250 [06:49<02:25,  2.21s/it]\n",
      "Loss: 0.01113:  74%|███████▍  | 185/250 [06:49<02:25,  2.25s/it]\n",
      "Loss: 0.01019:  74%|███████▍  | 185/250 [06:49<02:25,  2.25s/it]\n",
      "Loss: 0.01129:  74%|███████▍  | 185/250 [06:49<02:25,  2.25s/it]\n",
      "Loss: 0.00939:  74%|███████▍  | 185/250 [06:49<02:25,  2.25s/it]\n",
      "Loss: 0.00892:  74%|███████▍  | 185/250 [06:49<02:25,  2.25s/it]\n",
      "Loss: 0.00903:  74%|███████▍  | 185/250 [06:49<02:25,  2.25s/it]\n",
      "Loss: 0.01084:  74%|███████▍  | 185/250 [06:49<02:25,  2.25s/it]\n",
      "Loss: 0.01141:  74%|███████▍  | 185/250 [06:50<02:25,  2.25s/it]\n",
      "Loss: 0.01145:  74%|███████▍  | 185/250 [06:50<02:25,  2.25s/it]\n",
      "Loss: 0.01158:  74%|███████▍  | 185/250 [06:50<02:25,  2.25s/it]\n",
      "Loss: 0.01223:  74%|███████▍  | 185/250 [06:50<02:25,  2.25s/it]\n",
      "Loss: 0.01184:  74%|███████▍  | 185/250 [06:50<02:25,  2.25s/it]\n",
      "Loss: 0.01084:  74%|███████▍  | 185/250 [06:50<02:25,  2.25s/it]\n",
      "Loss: 0.01096:  74%|███████▍  | 185/250 [06:50<02:25,  2.25s/it]\n",
      "Loss: 0.01057:  74%|███████▍  | 185/250 [06:50<02:25,  2.25s/it]\n",
      "Loss: 0.00947:  74%|███████▍  | 185/250 [06:50<02:25,  2.25s/it]\n",
      "Loss: 0.00953:  74%|███████▍  | 185/250 [06:51<02:25,  2.25s/it]\n",
      "Loss: 0.00986:  74%|███████▍  | 185/250 [06:51<02:25,  2.25s/it]\n",
      "Loss: 0.01110:  74%|███████▍  | 185/250 [06:51<02:25,  2.25s/it]\n",
      "Loss: 0.01029:  74%|███████▍  | 185/250 [06:51<02:25,  2.25s/it]\n",
      "Loss: 0.01191:  74%|███████▍  | 185/250 [06:51<02:25,  2.25s/it]\n",
      "Loss: 0.01109:  74%|███████▍  | 185/250 [06:51<02:25,  2.25s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 55.85it/s]\u001b[A\n",
      "Loss: 0.01109:  74%|███████▍  | 186/250 [06:51<02:25,  2.27s/it]\n",
      "Loss: 0.01028:  74%|███████▍  | 186/250 [06:51<02:25,  2.27s/it]\n",
      "Loss: 0.01077:  74%|███████▍  | 186/250 [06:51<02:25,  2.27s/it]\n",
      "Loss: 0.01237:  74%|███████▍  | 186/250 [06:51<02:25,  2.27s/it]\n",
      "Loss: 0.01014:  74%|███████▍  | 186/250 [06:52<02:25,  2.27s/it]\n",
      "Loss: 0.00929:  74%|███████▍  | 186/250 [06:52<02:25,  2.27s/it]\n",
      "Loss: 0.01106:  74%|███████▍  | 186/250 [06:52<02:25,  2.27s/it]\n",
      "Loss: 0.01063:  74%|███████▍  | 186/250 [06:52<02:25,  2.27s/it]\n",
      "Loss: 0.01052:  74%|███████▍  | 186/250 [06:52<02:25,  2.27s/it]\n",
      "Loss: 0.01152:  74%|███████▍  | 186/250 [06:52<02:25,  2.27s/it]\n",
      "Loss: 0.01223:  74%|███████▍  | 186/250 [06:52<02:25,  2.27s/it]\n",
      "Loss: 0.01210:  74%|███████▍  | 186/250 [06:52<02:25,  2.27s/it]\n",
      "Loss: 0.01094:  74%|███████▍  | 186/250 [06:52<02:25,  2.27s/it]\n",
      "Loss: 0.01111:  74%|███████▍  | 186/250 [06:53<02:25,  2.27s/it]\n",
      "Loss: 0.01031:  74%|███████▍  | 186/250 [06:53<02:25,  2.27s/it]\n",
      "Loss: 0.00928:  74%|███████▍  | 186/250 [06:53<02:25,  2.27s/it]\n",
      "Loss: 0.01006:  74%|███████▍  | 186/250 [06:53<02:25,  2.27s/it]\n",
      "Loss: 0.00991:  74%|███████▍  | 186/250 [06:53<02:25,  2.27s/it]\n",
      "Loss: 0.01124:  74%|███████▍  | 186/250 [06:53<02:25,  2.27s/it]\n",
      "Loss: 0.01128:  74%|███████▍  | 186/250 [06:53<02:25,  2.27s/it]\n",
      "Loss: 0.01014:  74%|███████▍  | 186/250 [06:53<02:25,  2.27s/it]\n",
      "Loss: 0.01133:  74%|███████▍  | 186/250 [06:53<02:25,  2.27s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 54.53it/s]\u001b[A\n",
      "Loss: 0.01133:  75%|███████▍  | 187/250 [06:53<02:24,  2.29s/it]\n",
      "Loss: 0.01029:  75%|███████▍  | 187/250 [06:54<02:24,  2.29s/it]\n",
      "Loss: 0.01096:  75%|███████▍  | 187/250 [06:54<02:24,  2.29s/it]\n",
      "Loss: 0.00943:  75%|███████▍  | 187/250 [06:54<02:24,  2.29s/it]\n",
      "Loss: 0.01010:  75%|███████▍  | 187/250 [06:54<02:24,  2.29s/it]\n",
      "Loss: 0.01026:  75%|███████▍  | 187/250 [06:54<02:24,  2.29s/it]\n",
      "Loss: 0.01026:  75%|███████▍  | 187/250 [06:54<02:24,  2.29s/it]\n",
      "Loss: 0.01132:  75%|███████▍  | 187/250 [06:54<02:24,  2.29s/it]\n",
      "Loss: 0.01283:  75%|███████▍  | 187/250 [06:54<02:24,  2.29s/it]\n",
      "Loss: 0.01074:  75%|███████▍  | 187/250 [06:54<02:24,  2.29s/it]\n",
      "Loss: 0.01057:  75%|███████▍  | 187/250 [06:55<02:24,  2.29s/it]\n",
      "Loss: 0.01177:  75%|███████▍  | 187/250 [06:55<02:24,  2.29s/it]\n",
      "Loss: 0.01005:  75%|███████▍  | 187/250 [06:55<02:24,  2.29s/it]\n",
      "Loss: 0.01116:  75%|███████▍  | 187/250 [06:55<02:24,  2.29s/it]\n",
      "Loss: 0.01068:  75%|███████▍  | 187/250 [06:55<02:24,  2.29s/it]\n",
      "Loss: 0.00966:  75%|███████▍  | 187/250 [06:55<02:24,  2.29s/it]\n",
      "Loss: 0.01004:  75%|███████▍  | 187/250 [06:55<02:24,  2.29s/it]\n",
      "Loss: 0.01002:  75%|███████▍  | 187/250 [06:55<02:24,  2.29s/it]\n",
      "Loss: 0.01116:  75%|███████▍  | 187/250 [06:55<02:24,  2.29s/it]\n",
      "Loss: 0.01117:  75%|███████▍  | 187/250 [06:56<02:24,  2.29s/it]\n",
      "Loss: 0.00984:  75%|███████▍  | 187/250 [06:56<02:24,  2.29s/it]\n",
      "Loss: 0.01112:  75%|███████▍  | 187/250 [06:56<02:24,  2.29s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 54.63it/s]\u001b[A\n",
      "Loss: 0.01112:  75%|███████▌  | 188/250 [06:56<02:23,  2.32s/it]\n",
      "Loss: 0.01075:  75%|███████▌  | 188/250 [06:56<02:23,  2.32s/it]\n",
      "Loss: 0.01125:  75%|███████▌  | 188/250 [06:56<02:23,  2.32s/it]\n",
      "Loss: 0.00937:  75%|███████▌  | 188/250 [06:56<02:23,  2.32s/it]\n",
      "Loss: 0.01035:  75%|███████▌  | 188/250 [06:56<02:23,  2.32s/it]\n",
      "Loss: 0.01049:  75%|███████▌  | 188/250 [06:56<02:23,  2.32s/it]\n",
      "Loss: 0.00982:  75%|███████▌  | 188/250 [06:56<02:23,  2.32s/it]\n",
      "Loss: 0.01097:  75%|███████▌  | 188/250 [06:57<02:23,  2.32s/it]\n",
      "Loss: 0.01265:  75%|███████▌  | 188/250 [06:57<02:23,  2.32s/it]\n",
      "Loss: 0.01056:  75%|███████▌  | 188/250 [06:57<02:23,  2.32s/it]\n",
      "Loss: 0.01060:  75%|███████▌  | 188/250 [06:57<02:23,  2.32s/it]\n",
      "Loss: 0.01217:  75%|███████▌  | 188/250 [06:57<02:23,  2.32s/it]\n",
      "Loss: 0.01080:  75%|███████▌  | 188/250 [06:57<02:23,  2.32s/it]\n",
      "Loss: 0.01094:  75%|███████▌  | 188/250 [06:57<02:23,  2.32s/it]\n",
      "Loss: 0.01073:  75%|███████▌  | 188/250 [06:57<02:23,  2.32s/it]\n",
      "Loss: 0.00943:  75%|███████▌  | 188/250 [06:57<02:23,  2.32s/it]\n",
      "Loss: 0.00984:  75%|███████▌  | 188/250 [06:58<02:23,  2.32s/it]\n",
      "Loss: 0.00996:  75%|███████▌  | 188/250 [06:58<02:23,  2.32s/it]\n",
      "Loss: 0.01097:  75%|███████▌  | 188/250 [06:58<02:23,  2.32s/it]\n",
      "Loss: 0.01080:  75%|███████▌  | 188/250 [06:58<02:23,  2.32s/it]\n",
      "Loss: 0.01181:  75%|███████▌  | 188/250 [06:58<02:23,  2.32s/it]\n",
      "Loss: 0.01091:  75%|███████▌  | 188/250 [06:58<02:23,  2.32s/it]\n",
      "Loss: 0.01121:  75%|███████▌  | 188/250 [06:58<02:23,  2.32s/it]\n",
      "Loss: 0.01121:  76%|███████▌  | 189/250 [06:58<02:21,  2.32s/it]\n",
      "Loss: 0.01000:  76%|███████▌  | 189/250 [06:58<02:21,  2.32s/it]\n",
      "Loss: 0.01169:  76%|███████▌  | 189/250 [06:58<02:21,  2.32s/it]\n",
      "Loss: 0.01211:  76%|███████▌  | 189/250 [06:58<02:21,  2.32s/it]\n",
      "Loss: 0.00871:  76%|███████▌  | 189/250 [06:59<02:21,  2.32s/it]\n",
      "Loss: 0.00990:  76%|███████▌  | 189/250 [06:59<02:21,  2.32s/it]\n",
      "Loss: 0.01102:  76%|███████▌  | 189/250 [06:59<02:21,  2.32s/it]\n",
      "Loss: 0.01071:  76%|███████▌  | 189/250 [06:59<02:21,  2.32s/it]\n",
      "Loss: 0.01279:  76%|███████▌  | 189/250 [06:59<02:21,  2.32s/it]\n",
      "Loss: 0.01048:  76%|███████▌  | 189/250 [06:59<02:21,  2.32s/it]\n",
      "Loss: 0.01077:  76%|███████▌  | 189/250 [06:59<02:21,  2.32s/it]\n",
      "Loss: 0.01213:  76%|███████▌  | 189/250 [06:59<02:21,  2.32s/it]\n",
      "Loss: 0.01044:  76%|███████▌  | 189/250 [06:59<02:21,  2.32s/it]\n",
      "Loss: 0.01144:  76%|███████▌  | 189/250 [07:00<02:21,  2.32s/it]\n",
      "Loss: 0.01109:  76%|███████▌  | 189/250 [07:00<02:21,  2.32s/it]\n",
      "Loss: 0.01126:  76%|███████▌  | 189/250 [07:00<02:21,  2.32s/it]\n",
      "Loss: 0.01078:  76%|███████▌  | 189/250 [07:00<02:21,  2.32s/it]\n",
      "Loss: 0.01025:  76%|███████▌  | 189/250 [07:00<02:21,  2.32s/it]\n",
      "Loss: 0.01133:  76%|███████▌  | 189/250 [07:00<02:21,  2.32s/it]\n",
      "Loss: 0.01117:  76%|███████▌  | 189/250 [07:00<02:21,  2.32s/it]\n",
      "Loss: 0.00976:  76%|███████▌  | 189/250 [07:00<02:21,  2.32s/it]\n",
      "Loss: 0.01132:  76%|███████▌  | 189/250 [07:00<02:21,  2.32s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 52.80it/s]\u001b[A\n",
      "Loss: 0.01132:  76%|███████▌  | 190/250 [07:00<02:19,  2.33s/it]\n",
      "Loss: 0.01103:  76%|███████▌  | 190/250 [07:01<02:19,  2.33s/it]\n",
      "Loss: 0.01073:  76%|███████▌  | 190/250 [07:01<02:19,  2.33s/it]\n",
      "Loss: 0.01064:  76%|███████▌  | 190/250 [07:01<02:19,  2.33s/it]\n",
      "Loss: 0.00970:  76%|███████▌  | 190/250 [07:01<02:19,  2.33s/it]\n",
      "Loss: 0.00929:  76%|███████▌  | 190/250 [07:01<02:19,  2.33s/it]\n",
      "Loss: 0.01113:  76%|███████▌  | 190/250 [07:01<02:19,  2.33s/it]\n",
      "Loss: 0.00893:  76%|███████▌  | 190/250 [07:01<02:19,  2.33s/it]\n",
      "Loss: 0.01368:  76%|███████▌  | 190/250 [07:01<02:19,  2.33s/it]\n",
      "Loss: 0.00948:  76%|███████▌  | 190/250 [07:01<02:19,  2.33s/it]\n",
      "Loss: 0.01226:  76%|███████▌  | 190/250 [07:02<02:19,  2.33s/it]\n",
      "Loss: 0.01235:  76%|███████▌  | 190/250 [07:02<02:19,  2.33s/it]\n",
      "Loss: 0.01049:  76%|███████▌  | 190/250 [07:02<02:19,  2.33s/it]\n",
      "Loss: 0.01188:  76%|███████▌  | 190/250 [07:02<02:19,  2.33s/it]\n",
      "Loss: 0.01238:  76%|███████▌  | 190/250 [07:02<02:19,  2.33s/it]\n",
      "Loss: 0.01029:  76%|███████▌  | 190/250 [07:02<02:19,  2.33s/it]\n",
      "Loss: 0.01179:  76%|███████▌  | 190/250 [07:02<02:19,  2.33s/it]\n",
      "Loss: 0.01009:  76%|███████▌  | 190/250 [07:02<02:19,  2.33s/it]\n",
      "Loss: 0.01208:  76%|███████▌  | 190/250 [07:02<02:19,  2.33s/it]\n",
      "Loss: 0.01200:  76%|███████▌  | 190/250 [07:03<02:19,  2.33s/it]\n",
      "Loss: 0.01044:  76%|███████▌  | 190/250 [07:03<02:19,  2.33s/it]\n",
      "Loss: 0.01157:  76%|███████▌  | 190/250 [07:03<02:19,  2.33s/it]\n",
      "Loss: 0.01157:  76%|███████▋  | 191/250 [07:03<02:17,  2.32s/it]\n",
      "Loss: 0.01102:  76%|███████▋  | 191/250 [07:03<02:17,  2.32s/it]\n",
      "Loss: 0.01102:  76%|███████▋  | 191/250 [07:03<02:17,  2.32s/it]\n",
      "Loss: 0.00953:  76%|███████▋  | 191/250 [07:03<02:17,  2.32s/it]\n",
      "Loss: 0.01031:  76%|███████▋  | 191/250 [07:03<02:17,  2.32s/it]\n",
      "Loss: 0.01048:  76%|███████▋  | 191/250 [07:03<02:17,  2.32s/it]\n",
      "Loss: 0.00984:  76%|███████▋  | 191/250 [07:03<02:17,  2.32s/it]\n",
      "Loss: 0.01114:  76%|███████▋  | 191/250 [07:04<02:17,  2.32s/it]\n",
      "Loss: 0.01276:  76%|███████▋  | 191/250 [07:04<02:17,  2.32s/it]\n",
      "Loss: 0.01060:  76%|███████▋  | 191/250 [07:04<02:17,  2.32s/it]\n",
      "Loss: 0.01043:  76%|███████▋  | 191/250 [07:04<02:17,  2.32s/it]\n",
      "Loss: 0.01180:  76%|███████▋  | 191/250 [07:04<02:17,  2.32s/it]\n",
      "Loss: 0.01021:  76%|███████▋  | 191/250 [07:04<02:17,  2.32s/it]\n",
      "Loss: 0.01121:  76%|███████▋  | 191/250 [07:04<02:17,  2.32s/it]\n",
      "Loss: 0.01088:  76%|███████▋  | 191/250 [07:04<02:17,  2.32s/it]\n",
      "Loss: 0.01186:  76%|███████▋  | 191/250 [07:04<02:17,  2.32s/it]\n",
      "Loss: 0.01100:  76%|███████▋  | 191/250 [07:05<02:17,  2.32s/it]\n",
      "Loss: 0.01082:  76%|███████▋  | 191/250 [07:05<02:17,  2.32s/it]\n",
      "Loss: 0.01100:  76%|███████▋  | 191/250 [07:05<02:17,  2.32s/it]\n",
      "Loss: 0.01324:  76%|███████▋  | 191/250 [07:05<02:17,  2.32s/it]\n",
      "Loss: 0.01071:  76%|███████▋  | 191/250 [07:05<02:17,  2.32s/it]\n",
      "Loss: 0.01102:  76%|███████▋  | 191/250 [07:05<02:17,  2.32s/it]\n",
      "Loss: 0.01102:  77%|███████▋  | 192/250 [07:05<02:14,  2.32s/it]\n",
      "Loss: 0.01017:  77%|███████▋  | 192/250 [07:05<02:14,  2.32s/it]\n",
      "Loss: 0.01094:  77%|███████▋  | 192/250 [07:05<02:14,  2.32s/it]\n",
      "Loss: 0.00950:  77%|███████▋  | 192/250 [07:05<02:14,  2.32s/it]\n",
      "Loss: 0.01056:  77%|███████▋  | 192/250 [07:06<02:14,  2.32s/it]\n",
      "Loss: 0.00888:  77%|███████▋  | 192/250 [07:06<02:14,  2.32s/it]\n",
      "Loss: 0.01098:  77%|███████▋  | 192/250 [07:06<02:14,  2.32s/it]\n",
      "Loss: 0.01063:  77%|███████▋  | 192/250 [07:06<02:14,  2.32s/it]\n",
      "Loss: 0.01046:  77%|███████▋  | 192/250 [07:06<02:14,  2.32s/it]\n",
      "Loss: 0.01162:  77%|███████▋  | 192/250 [07:06<02:14,  2.32s/it]\n",
      "Loss: 0.01236:  77%|███████▋  | 192/250 [07:06<02:14,  2.32s/it]\n",
      "Loss: 0.01187:  77%|███████▋  | 192/250 [07:06<02:14,  2.32s/it]\n",
      "Loss: 0.01015:  77%|███████▋  | 192/250 [07:06<02:14,  2.32s/it]\n",
      "Loss: 0.01118:  77%|███████▋  | 192/250 [07:07<02:14,  2.32s/it]\n",
      "Loss: 0.01091:  77%|███████▋  | 192/250 [07:07<02:14,  2.32s/it]\n",
      "Loss: 0.01178:  77%|███████▋  | 192/250 [07:07<02:14,  2.32s/it]\n",
      "Loss: 0.01092:  77%|███████▋  | 192/250 [07:07<02:14,  2.32s/it]\n",
      "Loss: 0.01042:  77%|███████▋  | 192/250 [07:07<02:14,  2.32s/it]\n",
      "Loss: 0.01129:  77%|███████▋  | 192/250 [07:07<02:14,  2.32s/it]\n",
      "Loss: 0.01168:  77%|███████▋  | 192/250 [07:07<02:14,  2.32s/it]\n",
      "Loss: 0.00978:  77%|███████▋  | 192/250 [07:07<02:14,  2.32s/it]\n",
      "Loss: 0.01134:  77%|███████▋  | 192/250 [07:07<02:14,  2.32s/it]\n",
      "Loss: 0.01134:  77%|███████▋  | 193/250 [07:07<02:11,  2.31s/it]\n",
      "Loss: 0.01004:  77%|███████▋  | 193/250 [07:07<02:11,  2.31s/it]\n",
      "Loss: 0.01078:  77%|███████▋  | 193/250 [07:08<02:11,  2.31s/it]\n",
      "Loss: 0.01267:  77%|███████▋  | 193/250 [07:08<02:11,  2.31s/it]\n",
      "Loss: 0.00951:  77%|███████▋  | 193/250 [07:08<02:11,  2.31s/it]\n",
      "Loss: 0.01045:  77%|███████▋  | 193/250 [07:08<02:11,  2.31s/it]\n",
      "Loss: 0.01140:  77%|███████▋  | 193/250 [07:08<02:11,  2.31s/it]\n",
      "Loss: 0.00886:  77%|███████▋  | 193/250 [07:08<02:11,  2.31s/it]\n",
      "Loss: 0.01376:  77%|███████▋  | 193/250 [07:08<02:11,  2.31s/it]\n",
      "Loss: 0.00967:  77%|███████▋  | 193/250 [07:08<02:11,  2.31s/it]\n",
      "Loss: 0.01199:  77%|███████▋  | 193/250 [07:08<02:11,  2.31s/it]\n",
      "Loss: 0.01269:  77%|███████▋  | 193/250 [07:09<02:11,  2.31s/it]\n",
      "Loss: 0.01046:  77%|███████▋  | 193/250 [07:09<02:11,  2.31s/it]\n",
      "Loss: 0.01165:  77%|███████▋  | 193/250 [07:09<02:11,  2.31s/it]\n",
      "Loss: 0.01227:  77%|███████▋  | 193/250 [07:09<02:11,  2.31s/it]\n",
      "Loss: 0.00974:  77%|███████▋  | 193/250 [07:09<02:11,  2.31s/it]\n",
      "Loss: 0.01020:  77%|███████▋  | 193/250 [07:09<02:11,  2.31s/it]\n",
      "Loss: 0.01021:  77%|███████▋  | 193/250 [07:09<02:11,  2.31s/it]\n",
      "Loss: 0.01226:  77%|███████▋  | 193/250 [07:09<02:11,  2.31s/it]\n",
      "Loss: 0.01193:  77%|███████▋  | 193/250 [07:09<02:11,  2.31s/it]\n",
      "Loss: 0.01037:  77%|███████▋  | 193/250 [07:10<02:11,  2.31s/it]\n",
      "Loss: 0.01153:  77%|███████▋  | 193/250 [07:10<02:11,  2.31s/it]\n",
      "Loss: 0.01153:  78%|███████▊  | 194/250 [07:10<02:08,  2.29s/it]\n",
      "Loss: 0.01060:  78%|███████▊  | 194/250 [07:10<02:08,  2.29s/it]\n",
      "Loss: 0.01105:  78%|███████▊  | 194/250 [07:10<02:08,  2.29s/it]\n",
      "Loss: 0.01263:  78%|███████▊  | 194/250 [07:10<02:08,  2.29s/it]\n",
      "Loss: 0.00958:  78%|███████▊  | 194/250 [07:10<02:08,  2.29s/it]\n",
      "Loss: 0.00967:  78%|███████▊  | 194/250 [07:10<02:08,  2.29s/it]\n",
      "Loss: 0.01115:  78%|███████▊  | 194/250 [07:10<02:08,  2.29s/it]\n",
      "Loss: 0.00888:  78%|███████▊  | 194/250 [07:10<02:08,  2.29s/it]\n",
      "Loss: 0.01363:  78%|███████▊  | 194/250 [07:10<02:08,  2.29s/it]\n",
      "Loss: 0.00987:  78%|███████▊  | 194/250 [07:11<02:08,  2.29s/it]\n",
      "Loss: 0.01213:  78%|███████▊  | 194/250 [07:11<02:08,  2.29s/it]\n",
      "Loss: 0.01091:  78%|███████▊  | 194/250 [07:11<02:08,  2.29s/it]\n",
      "Loss: 0.01011:  78%|███████▊  | 194/250 [07:11<02:08,  2.29s/it]\n",
      "Loss: 0.01088:  78%|███████▊  | 194/250 [07:11<02:08,  2.29s/it]\n",
      "Loss: 0.00950:  78%|███████▊  | 194/250 [07:11<02:08,  2.29s/it]\n",
      "Loss: 0.01009:  78%|███████▊  | 194/250 [07:11<02:08,  2.29s/it]\n",
      "Loss: 0.01210:  78%|███████▊  | 194/250 [07:11<02:08,  2.29s/it]\n",
      "Loss: 0.00989:  78%|███████▊  | 194/250 [07:11<02:08,  2.29s/it]\n",
      "Loss: 0.01207:  78%|███████▊  | 194/250 [07:12<02:08,  2.29s/it]\n",
      "Loss: 0.01173:  78%|███████▊  | 194/250 [07:12<02:08,  2.29s/it]\n",
      "Loss: 0.01058:  78%|███████▊  | 194/250 [07:12<02:08,  2.29s/it]\n",
      "Loss: 0.01123:  78%|███████▊  | 194/250 [07:12<02:08,  2.29s/it]\n",
      "Loss: 0.01123:  78%|███████▊  | 195/250 [07:12<02:04,  2.27s/it]\n",
      "Loss: 0.01033:  78%|███████▊  | 195/250 [07:12<02:04,  2.27s/it]\n",
      "Loss: 0.01085:  78%|███████▊  | 195/250 [07:12<02:04,  2.27s/it]\n",
      "Loss: 0.01266:  78%|███████▊  | 195/250 [07:12<02:04,  2.27s/it]\n",
      "Loss: 0.01205:  78%|███████▊  | 195/250 [07:12<02:04,  2.27s/it]\n",
      "Loss: 0.01203:  78%|███████▊  | 195/250 [07:12<02:04,  2.27s/it]\n",
      "Loss: 0.01093:  78%|███████▊  | 195/250 [07:12<02:04,  2.27s/it]\n",
      "Loss: 0.00957:  78%|███████▊  | 195/250 [07:13<02:04,  2.27s/it]\n",
      "Loss: 0.01116:  78%|███████▊  | 195/250 [07:13<02:04,  2.27s/it]\n",
      "Loss: 0.01046:  78%|███████▊  | 195/250 [07:13<02:04,  2.27s/it]\n",
      "Loss: 0.01191:  78%|███████▊  | 195/250 [07:13<02:04,  2.27s/it]\n",
      "Loss: 0.01097:  78%|███████▊  | 195/250 [07:13<02:04,  2.27s/it]\n",
      "Loss: 0.01152:  78%|███████▊  | 195/250 [07:13<02:04,  2.27s/it]\n",
      "Loss: 0.01172:  78%|███████▊  | 195/250 [07:13<02:04,  2.27s/it]\n",
      "Loss: 0.01128:  78%|███████▊  | 195/250 [07:13<02:04,  2.27s/it]\n",
      "Loss: 0.01129:  78%|███████▊  | 195/250 [07:13<02:04,  2.27s/it]\n",
      "Loss: 0.01073:  78%|███████▊  | 195/250 [07:14<02:04,  2.27s/it]\n",
      "Loss: 0.01102:  78%|███████▊  | 195/250 [07:14<02:04,  2.27s/it]\n",
      "Loss: 0.01064:  78%|███████▊  | 195/250 [07:14<02:04,  2.27s/it]\n",
      "Loss: 0.00990:  78%|███████▊  | 195/250 [07:14<02:04,  2.27s/it]\n",
      "Loss: 0.01055:  78%|███████▊  | 195/250 [07:14<02:04,  2.27s/it]\n",
      "Loss: 0.01097:  78%|███████▊  | 195/250 [07:14<02:04,  2.27s/it]\n",
      "Loss: 0.01097:  78%|███████▊  | 196/250 [07:14<02:01,  2.24s/it]\n",
      "Loss: 0.01025:  78%|███████▊  | 196/250 [07:14<02:01,  2.24s/it]\n",
      "Loss: 0.01069:  78%|███████▊  | 196/250 [07:14<02:01,  2.24s/it]\n",
      "Loss: 0.01255:  78%|███████▊  | 196/250 [07:14<02:01,  2.24s/it]\n",
      "Loss: 0.00914:  78%|███████▊  | 196/250 [07:14<02:01,  2.24s/it]\n",
      "Loss: 0.00916:  78%|███████▊  | 196/250 [07:15<02:01,  2.24s/it]\n",
      "Loss: 0.01100:  78%|███████▊  | 196/250 [07:15<02:01,  2.24s/it]\n",
      "Loss: 0.00898:  78%|███████▊  | 196/250 [07:15<02:01,  2.24s/it]\n",
      "Loss: 0.01365:  78%|███████▊  | 196/250 [07:15<02:01,  2.24s/it]\n",
      "Loss: 0.00931:  78%|███████▊  | 196/250 [07:15<02:01,  2.24s/it]\n",
      "Loss: 0.01212:  78%|███████▊  | 196/250 [07:15<02:01,  2.24s/it]\n",
      "Loss: 0.01271:  78%|███████▊  | 196/250 [07:15<02:01,  2.24s/it]\n",
      "Loss: 0.01028:  78%|███████▊  | 196/250 [07:15<02:01,  2.24s/it]\n",
      "Loss: 0.01202:  78%|███████▊  | 196/250 [07:15<02:01,  2.24s/it]\n",
      "Loss: 0.01244:  78%|███████▊  | 196/250 [07:16<02:01,  2.24s/it]\n",
      "Loss: 0.01067:  78%|███████▊  | 196/250 [07:16<02:01,  2.24s/it]\n",
      "Loss: 0.00997:  78%|███████▊  | 196/250 [07:16<02:01,  2.24s/it]\n",
      "Loss: 0.01002:  78%|███████▊  | 196/250 [07:16<02:01,  2.24s/it]\n",
      "Loss: 0.01225:  78%|███████▊  | 196/250 [07:16<02:01,  2.24s/it]\n",
      "Loss: 0.01216:  78%|███████▊  | 196/250 [07:16<02:01,  2.24s/it]\n",
      "Loss: 0.01038:  78%|███████▊  | 196/250 [07:16<02:01,  2.24s/it]\n",
      "Loss: 0.01169:  78%|███████▊  | 196/250 [07:16<02:01,  2.24s/it]\n",
      "Loss: 0.01169:  79%|███████▉  | 197/250 [07:16<01:58,  2.24s/it]\n",
      "Loss: 0.01090:  79%|███████▉  | 197/250 [07:16<01:58,  2.24s/it]\n",
      "Loss: 0.01092:  79%|███████▉  | 197/250 [07:16<01:58,  2.24s/it]\n",
      "Loss: 0.01098:  79%|███████▉  | 197/250 [07:17<01:58,  2.24s/it]\n",
      "Loss: 0.01221:  79%|███████▉  | 197/250 [07:17<01:58,  2.24s/it]\n",
      "Loss: 0.01195:  79%|███████▉  | 197/250 [07:17<01:58,  2.24s/it]\n",
      "Loss: 0.01080:  79%|███████▉  | 197/250 [07:17<01:58,  2.24s/it]\n",
      "Loss: 0.01192:  79%|███████▉  | 197/250 [07:17<01:58,  2.24s/it]\n",
      "Loss: 0.01156:  79%|███████▉  | 197/250 [07:17<01:58,  2.24s/it]\n",
      "Loss: 0.01033:  79%|███████▉  | 197/250 [07:17<01:58,  2.24s/it]\n",
      "Loss: 0.01197:  79%|███████▉  | 197/250 [07:17<01:58,  2.24s/it]\n",
      "Loss: 0.01134:  79%|███████▉  | 197/250 [07:17<01:58,  2.24s/it]\n",
      "Loss: 0.01148:  79%|███████▉  | 197/250 [07:18<01:58,  2.24s/it]\n",
      "Loss: 0.01183:  79%|███████▉  | 197/250 [07:18<01:58,  2.24s/it]\n",
      "Loss: 0.01114:  79%|███████▉  | 197/250 [07:18<01:58,  2.24s/it]\n",
      "Loss: 0.01144:  79%|███████▉  | 197/250 [07:18<01:58,  2.24s/it]\n",
      "Loss: 0.01053:  79%|███████▉  | 197/250 [07:18<01:58,  2.24s/it]\n",
      "Loss: 0.01133:  79%|███████▉  | 197/250 [07:18<01:58,  2.24s/it]\n",
      "Loss: 0.01025:  79%|███████▉  | 197/250 [07:18<01:58,  2.24s/it]\n",
      "Loss: 0.01097:  79%|███████▉  | 197/250 [07:18<01:58,  2.24s/it]\n",
      "Loss: 0.01075:  79%|███████▉  | 197/250 [07:18<01:58,  2.24s/it]\n",
      "Loss: 0.01105:  79%|███████▉  | 197/250 [07:18<01:58,  2.24s/it]\n",
      "Loss: 0.01105:  79%|███████▉  | 198/250 [07:18<01:55,  2.22s/it]\n",
      "Loss: 0.01062:  79%|███████▉  | 198/250 [07:19<01:55,  2.22s/it]\n",
      "Loss: 0.01077:  79%|███████▉  | 198/250 [07:19<01:55,  2.22s/it]\n",
      "Loss: 0.01073:  79%|███████▉  | 198/250 [07:19<01:55,  2.22s/it]\n",
      "Loss: 0.01215:  79%|███████▉  | 198/250 [07:19<01:55,  2.22s/it]\n",
      "Loss: 0.01201:  79%|███████▉  | 198/250 [07:19<01:55,  2.22s/it]\n",
      "Loss: 0.01101:  79%|███████▉  | 198/250 [07:19<01:55,  2.22s/it]\n",
      "Loss: 0.01167:  79%|███████▉  | 198/250 [07:19<01:55,  2.22s/it]\n",
      "Loss: 0.01033:  79%|███████▉  | 198/250 [07:19<01:55,  2.22s/it]\n",
      "Loss: 0.01244:  79%|███████▉  | 198/250 [07:19<01:55,  2.22s/it]\n",
      "Loss: 0.01145:  79%|███████▉  | 198/250 [07:20<01:55,  2.22s/it]\n",
      "Loss: 0.01018:  79%|███████▉  | 198/250 [07:20<01:55,  2.22s/it]\n",
      "Loss: 0.01043:  79%|███████▉  | 198/250 [07:20<01:55,  2.22s/it]\n",
      "Loss: 0.01077:  79%|███████▉  | 198/250 [07:20<01:55,  2.22s/it]\n",
      "Loss: 0.00962:  79%|███████▉  | 198/250 [07:20<01:55,  2.22s/it]\n",
      "Loss: 0.01016:  79%|███████▉  | 198/250 [07:20<01:55,  2.22s/it]\n",
      "Loss: 0.01191:  79%|███████▉  | 198/250 [07:20<01:55,  2.22s/it]\n",
      "Loss: 0.01042:  79%|███████▉  | 198/250 [07:20<01:55,  2.22s/it]\n",
      "Loss: 0.01227:  79%|███████▉  | 198/250 [07:20<01:55,  2.22s/it]\n",
      "Loss: 0.01184:  79%|███████▉  | 198/250 [07:20<01:55,  2.22s/it]\n",
      "Loss: 0.01039:  79%|███████▉  | 198/250 [07:21<01:55,  2.22s/it]\n",
      "Loss: 0.01116:  79%|███████▉  | 198/250 [07:21<01:55,  2.22s/it]\n",
      "Loss: 0.01116:  80%|███████▉  | 199/250 [07:21<01:52,  2.21s/it]\n",
      "Loss: 0.01025:  80%|███████▉  | 199/250 [07:21<01:52,  2.21s/it]\n",
      "Loss: 0.01093:  80%|███████▉  | 199/250 [07:21<01:52,  2.21s/it]\n",
      "Loss: 0.01262:  80%|███████▉  | 199/250 [07:21<01:52,  2.21s/it]\n",
      "Loss: 0.00913:  80%|███████▉  | 199/250 [07:21<01:52,  2.21s/it]\n",
      "Loss: 0.00928:  80%|███████▉  | 199/250 [07:21<01:52,  2.21s/it]\n",
      "Loss: 0.01130:  80%|███████▉  | 199/250 [07:21<01:52,  2.21s/it]\n",
      "Loss: 0.00877:  80%|███████▉  | 199/250 [07:21<01:52,  2.21s/it]\n",
      "Loss: 0.01354:  80%|███████▉  | 199/250 [07:21<01:52,  2.21s/it]\n",
      "Loss: 0.01224:  80%|███████▉  | 199/250 [07:22<01:52,  2.21s/it]\n",
      "Loss: 0.01125:  80%|███████▉  | 199/250 [07:22<01:52,  2.21s/it]\n",
      "Loss: 0.01012:  80%|███████▉  | 199/250 [07:22<01:52,  2.21s/it]\n",
      "Loss: 0.01030:  80%|███████▉  | 199/250 [07:22<01:52,  2.21s/it]\n",
      "Loss: 0.01067:  80%|███████▉  | 199/250 [07:22<01:52,  2.21s/it]\n",
      "Loss: 0.00962:  80%|███████▉  | 199/250 [07:22<01:52,  2.21s/it]\n",
      "Loss: 0.00990:  80%|███████▉  | 199/250 [07:22<01:52,  2.21s/it]\n",
      "Loss: 0.01161:  80%|███████▉  | 199/250 [07:22<01:52,  2.21s/it]\n",
      "Loss: 0.00982:  80%|███████▉  | 199/250 [07:22<01:52,  2.21s/it]\n",
      "Loss: 0.01008:  80%|███████▉  | 199/250 [07:23<01:52,  2.21s/it]\n",
      "Loss: 0.01094:  80%|███████▉  | 199/250 [07:23<01:52,  2.21s/it]\n",
      "Loss: 0.01012:  80%|███████▉  | 199/250 [07:23<01:52,  2.21s/it]\n",
      "Loss: 0.01100:  80%|███████▉  | 199/250 [07:23<01:52,  2.21s/it]\n",
      "Loss: 0.01100:  80%|████████  | 200/250 [07:23<01:50,  2.21s/it]\n",
      "Loss: 0.01029:  80%|████████  | 200/250 [07:23<01:50,  2.21s/it]\n",
      "Loss: 0.01094:  80%|████████  | 200/250 [07:23<01:50,  2.21s/it]\n",
      "Loss: 0.01213:  80%|████████  | 200/250 [07:23<01:50,  2.21s/it]\n",
      "Loss: 0.00900:  80%|████████  | 200/250 [07:23<01:50,  2.21s/it]\n",
      "Loss: 0.00941:  80%|████████  | 200/250 [07:23<01:50,  2.21s/it]\n",
      "Loss: 0.01129:  80%|████████  | 200/250 [07:23<01:50,  2.21s/it]\n",
      "Loss: 0.00884:  80%|████████  | 200/250 [07:24<01:50,  2.21s/it]\n",
      "Loss: 0.01373:  80%|████████  | 200/250 [07:24<01:50,  2.21s/it]\n",
      "Loss: 0.00948:  80%|████████  | 200/250 [07:24<01:50,  2.21s/it]\n",
      "Loss: 0.01228:  80%|████████  | 200/250 [07:24<01:50,  2.21s/it]\n",
      "Loss: 0.01224:  80%|████████  | 200/250 [07:24<01:50,  2.21s/it]\n",
      "Loss: 0.01045:  80%|████████  | 200/250 [07:24<01:50,  2.21s/it]\n",
      "Loss: 0.01070:  80%|████████  | 200/250 [07:24<01:50,  2.21s/it]\n",
      "Loss: 0.00946:  80%|████████  | 200/250 [07:24<01:50,  2.21s/it]\n",
      "Loss: 0.00992:  80%|████████  | 200/250 [07:24<01:50,  2.21s/it]\n",
      "Loss: 0.01157:  80%|████████  | 200/250 [07:25<01:50,  2.21s/it]\n",
      "Loss: 0.00985:  80%|████████  | 200/250 [07:25<01:50,  2.21s/it]\n",
      "Loss: 0.01207:  80%|████████  | 200/250 [07:25<01:50,  2.21s/it]\n",
      "Loss: 0.01169:  80%|████████  | 200/250 [07:25<01:50,  2.21s/it]\n",
      "Loss: 0.01051:  80%|████████  | 200/250 [07:25<01:50,  2.21s/it]\n",
      "Loss: 0.01096:  80%|████████  | 200/250 [07:25<01:50,  2.21s/it]\n",
      "Loss: 0.01096:  80%|████████  | 201/250 [07:25<01:48,  2.21s/it]\n",
      "Loss: 0.00999:  80%|████████  | 201/250 [07:25<01:48,  2.21s/it]\n",
      "Loss: 0.01098:  80%|████████  | 201/250 [07:25<01:48,  2.21s/it]\n",
      "Loss: 0.01232:  80%|████████  | 201/250 [07:25<01:48,  2.21s/it]\n",
      "Loss: 0.00935:  80%|████████  | 201/250 [07:25<01:48,  2.21s/it]\n",
      "Loss: 0.00966:  80%|████████  | 201/250 [07:26<01:48,  2.21s/it]\n",
      "Loss: 0.01121:  80%|████████  | 201/250 [07:26<01:48,  2.21s/it]\n",
      "Loss: 0.00865:  80%|████████  | 201/250 [07:26<01:48,  2.21s/it]\n",
      "Loss: 0.01017:  80%|████████  | 201/250 [07:26<01:48,  2.21s/it]\n",
      "Loss: 0.01251:  80%|████████  | 201/250 [07:26<01:48,  2.21s/it]\n",
      "Loss: 0.01104:  80%|████████  | 201/250 [07:26<01:48,  2.21s/it]\n",
      "Loss: 0.01005:  80%|████████  | 201/250 [07:26<01:48,  2.21s/it]\n",
      "Loss: 0.01055:  80%|████████  | 201/250 [07:26<01:48,  2.21s/it]\n",
      "Loss: 0.01044:  80%|████████  | 201/250 [07:26<01:48,  2.21s/it]\n",
      "Loss: 0.00946:  80%|████████  | 201/250 [07:27<01:48,  2.21s/it]\n",
      "Loss: 0.01026:  80%|████████  | 201/250 [07:27<01:48,  2.21s/it]\n",
      "Loss: 0.01192:  80%|████████  | 201/250 [07:27<01:48,  2.21s/it]\n",
      "Loss: 0.00989:  80%|████████  | 201/250 [07:27<01:48,  2.21s/it]\n",
      "Loss: 0.01193:  80%|████████  | 201/250 [07:27<01:48,  2.21s/it]\n",
      "Loss: 0.01175:  80%|████████  | 201/250 [07:27<01:48,  2.21s/it]\n",
      "Loss: 0.01058:  80%|████████  | 201/250 [07:27<01:48,  2.21s/it]\n",
      "Loss: 0.01112:  80%|████████  | 201/250 [07:27<01:48,  2.21s/it]\n",
      "Loss: 0.01112:  81%|████████  | 202/250 [07:27<01:46,  2.22s/it]\n",
      "Loss: 0.01069:  81%|████████  | 202/250 [07:27<01:46,  2.22s/it]\n",
      "Loss: 0.01062:  81%|████████  | 202/250 [07:28<01:46,  2.22s/it]\n",
      "Loss: 0.01075:  81%|████████  | 202/250 [07:28<01:46,  2.22s/it]\n",
      "Loss: 0.01178:  81%|████████  | 202/250 [07:28<01:46,  2.22s/it]\n",
      "Loss: 0.01192:  81%|████████  | 202/250 [07:28<01:46,  2.22s/it]\n",
      "Loss: 0.01078:  81%|████████  | 202/250 [07:28<01:46,  2.22s/it]\n",
      "Loss: 0.01165:  81%|████████  | 202/250 [07:28<01:46,  2.22s/it]\n",
      "Loss: 0.01072:  81%|████████  | 202/250 [07:28<01:46,  2.22s/it]\n",
      "Loss: 0.01018:  81%|████████  | 202/250 [07:28<01:46,  2.22s/it]\n",
      "Loss: 0.01155:  81%|████████  | 202/250 [07:28<01:46,  2.22s/it]\n",
      "Loss: 0.01109:  81%|████████  | 202/250 [07:28<01:46,  2.22s/it]\n",
      "Loss: 0.01123:  81%|████████  | 202/250 [07:29<01:46,  2.22s/it]\n",
      "Loss: 0.01204:  81%|████████  | 202/250 [07:29<01:46,  2.22s/it]\n",
      "Loss: 0.01113:  81%|████████  | 202/250 [07:29<01:46,  2.22s/it]\n",
      "Loss: 0.01105:  81%|████████  | 202/250 [07:29<01:46,  2.22s/it]\n",
      "Loss: 0.01057:  81%|████████  | 202/250 [07:29<01:46,  2.22s/it]\n",
      "Loss: 0.01093:  81%|████████  | 202/250 [07:29<01:46,  2.22s/it]\n",
      "Loss: 0.01061:  81%|████████  | 202/250 [07:29<01:46,  2.22s/it]\n",
      "Loss: 0.00962:  81%|████████  | 202/250 [07:29<01:46,  2.22s/it]\n",
      "Loss: 0.01024:  81%|████████  | 202/250 [07:29<01:46,  2.22s/it]\n",
      "Loss: 0.01102:  81%|████████  | 202/250 [07:29<01:46,  2.22s/it]\n",
      "Loss: 0.01102:  81%|████████  | 203/250 [07:29<01:43,  2.21s/it]\n",
      "Loss: 0.01029:  81%|████████  | 203/250 [07:30<01:43,  2.21s/it]\n",
      "Loss: 0.01092:  81%|████████  | 203/250 [07:30<01:43,  2.21s/it]\n",
      "Loss: 0.01255:  81%|████████  | 203/250 [07:30<01:43,  2.21s/it]\n",
      "Loss: 0.00938:  81%|████████  | 203/250 [07:30<01:43,  2.21s/it]\n",
      "Loss: 0.00942:  81%|████████  | 203/250 [07:30<01:43,  2.21s/it]\n",
      "Loss: 0.01140:  81%|████████  | 203/250 [07:30<01:43,  2.21s/it]\n",
      "Loss: 0.00875:  81%|████████  | 203/250 [07:30<01:43,  2.21s/it]\n",
      "Loss: 0.01372:  81%|████████  | 203/250 [07:30<01:43,  2.21s/it]\n",
      "Loss: 0.00955:  81%|████████  | 203/250 [07:30<01:43,  2.21s/it]\n",
      "Loss: 0.01190:  81%|████████  | 203/250 [07:31<01:43,  2.21s/it]\n",
      "Loss: 0.01208:  81%|████████  | 203/250 [07:31<01:43,  2.21s/it]\n",
      "Loss: 0.01050:  81%|████████  | 203/250 [07:31<01:43,  2.21s/it]\n",
      "Loss: 0.01130:  81%|████████  | 203/250 [07:31<01:43,  2.21s/it]\n",
      "Loss: 0.01214:  81%|████████  | 203/250 [07:31<01:43,  2.21s/it]\n",
      "Loss: 0.01017:  81%|████████  | 203/250 [07:31<01:43,  2.21s/it]\n",
      "Loss: 0.00970:  81%|████████  | 203/250 [07:31<01:43,  2.21s/it]\n",
      "Loss: 0.01079:  81%|████████  | 203/250 [07:31<01:43,  2.21s/it]\n",
      "Loss: 0.01116:  81%|████████  | 203/250 [07:31<01:43,  2.21s/it]\n",
      "Loss: 0.01338:  81%|████████  | 203/250 [07:31<01:43,  2.21s/it]\n",
      "Loss: 0.01080:  81%|████████  | 203/250 [07:32<01:43,  2.21s/it]\n",
      "Loss: 0.01101:  81%|████████  | 203/250 [07:32<01:43,  2.21s/it]\n",
      "Loss: 0.01101:  82%|████████▏ | 204/250 [07:32<01:41,  2.21s/it]\n",
      "Loss: 0.01011:  82%|████████▏ | 204/250 [07:32<01:41,  2.21s/it]\n",
      "Loss: 0.01067:  82%|████████▏ | 204/250 [07:32<01:41,  2.21s/it]\n",
      "Loss: 0.01269:  82%|████████▏ | 204/250 [07:32<01:41,  2.21s/it]\n",
      "Loss: 0.00934:  82%|████████▏ | 204/250 [07:32<01:41,  2.21s/it]\n",
      "Loss: 0.00912:  82%|████████▏ | 204/250 [07:32<01:41,  2.21s/it]\n",
      "Loss: 0.01062:  82%|████████▏ | 204/250 [07:32<01:41,  2.21s/it]\n",
      "Loss: 0.01152:  82%|████████▏ | 204/250 [07:32<01:41,  2.21s/it]\n",
      "Loss: 0.01060:  82%|████████▏ | 204/250 [07:33<01:41,  2.21s/it]\n",
      "Loss: 0.01228:  82%|████████▏ | 204/250 [07:33<01:41,  2.21s/it]\n",
      "Loss: 0.01111:  82%|████████▏ | 204/250 [07:33<01:41,  2.21s/it]\n",
      "Loss: 0.01009:  82%|████████▏ | 204/250 [07:33<01:41,  2.21s/it]\n",
      "Loss: 0.01100:  82%|████████▏ | 204/250 [07:33<01:41,  2.21s/it]\n",
      "Loss: 0.01058:  82%|████████▏ | 204/250 [07:33<01:41,  2.21s/it]\n",
      "Loss: 0.00992:  82%|████████▏ | 204/250 [07:33<01:41,  2.21s/it]\n",
      "Loss: 0.00988:  82%|████████▏ | 204/250 [07:33<01:41,  2.21s/it]\n",
      "Loss: 0.01152:  82%|████████▏ | 204/250 [07:33<01:41,  2.21s/it]\n",
      "Loss: 0.01025:  82%|████████▏ | 204/250 [07:33<01:41,  2.21s/it]\n",
      "Loss: 0.01194:  82%|████████▏ | 204/250 [07:34<01:41,  2.21s/it]\n",
      "Loss: 0.01141:  82%|████████▏ | 204/250 [07:34<01:41,  2.21s/it]\n",
      "Loss: 0.01054:  82%|████████▏ | 204/250 [07:34<01:41,  2.21s/it]\n",
      "Loss: 0.01121:  82%|████████▏ | 204/250 [07:34<01:41,  2.21s/it]\n",
      "Loss: 0.01121:  82%|████████▏ | 205/250 [07:34<01:39,  2.21s/it]\n",
      "Loss: 0.01028:  82%|████████▏ | 205/250 [07:34<01:39,  2.21s/it]\n",
      "Loss: 0.01066:  82%|████████▏ | 205/250 [07:34<01:39,  2.21s/it]\n",
      "Loss: 0.01219:  82%|████████▏ | 205/250 [07:34<01:39,  2.21s/it]\n",
      "Loss: 0.00924:  82%|████████▏ | 205/250 [07:34<01:39,  2.21s/it]\n",
      "Loss: 0.00937:  82%|████████▏ | 205/250 [07:34<01:39,  2.21s/it]\n",
      "Loss: 0.01111:  82%|████████▏ | 205/250 [07:35<01:39,  2.21s/it]\n",
      "Loss: 0.00859:  82%|████████▏ | 205/250 [07:35<01:39,  2.21s/it]\n",
      "Loss: 0.01367:  82%|████████▏ | 205/250 [07:35<01:39,  2.21s/it]\n",
      "Loss: 0.00938:  82%|████████▏ | 205/250 [07:35<01:39,  2.21s/it]\n",
      "Loss: 0.01185:  82%|████████▏ | 205/250 [07:35<01:39,  2.21s/it]\n",
      "Loss: 0.01247:  82%|████████▏ | 205/250 [07:35<01:39,  2.21s/it]\n",
      "Loss: 0.01015:  82%|████████▏ | 205/250 [07:35<01:39,  2.21s/it]\n",
      "Loss: 0.01138:  82%|████████▏ | 205/250 [07:35<01:39,  2.21s/it]\n",
      "Loss: 0.01219:  82%|████████▏ | 205/250 [07:35<01:39,  2.21s/it]\n",
      "Loss: 0.01031:  82%|████████▏ | 205/250 [07:35<01:39,  2.21s/it]\n",
      "Loss: 0.00962:  82%|████████▏ | 205/250 [07:36<01:39,  2.21s/it]\n",
      "Loss: 0.01064:  82%|████████▏ | 205/250 [07:36<01:39,  2.21s/it]\n",
      "Loss: 0.01080:  82%|████████▏ | 205/250 [07:36<01:39,  2.21s/it]\n",
      "Loss: 0.01305:  82%|████████▏ | 205/250 [07:36<01:39,  2.21s/it]\n",
      "Loss: 0.01039:  82%|████████▏ | 205/250 [07:36<01:39,  2.21s/it]\n",
      "Loss: 0.01099:  82%|████████▏ | 205/250 [07:36<01:39,  2.21s/it]\n",
      "Loss: 0.01099:  82%|████████▏ | 206/250 [07:36<01:37,  2.21s/it]\n",
      "Loss: 0.01003:  82%|████████▏ | 206/250 [07:36<01:37,  2.21s/it]\n",
      "Loss: 0.01068:  82%|████████▏ | 206/250 [07:36<01:37,  2.21s/it]\n",
      "Loss: 0.01231:  82%|████████▏ | 206/250 [07:36<01:37,  2.21s/it]\n",
      "Loss: 0.00931:  82%|████████▏ | 206/250 [07:37<01:37,  2.21s/it]\n",
      "Loss: 0.00903:  82%|████████▏ | 206/250 [07:37<01:37,  2.21s/it]\n",
      "Loss: 0.01130:  82%|████████▏ | 206/250 [07:37<01:37,  2.21s/it]\n",
      "Loss: 0.00852:  82%|████████▏ | 206/250 [07:37<01:37,  2.21s/it]\n",
      "Loss: 0.01365:  82%|████████▏ | 206/250 [07:37<01:37,  2.21s/it]\n",
      "Loss: 0.00961:  82%|████████▏ | 206/250 [07:37<01:37,  2.21s/it]\n",
      "Loss: 0.01212:  82%|████████▏ | 206/250 [07:37<01:37,  2.21s/it]\n",
      "Loss: 0.01237:  82%|████████▏ | 206/250 [07:37<01:37,  2.21s/it]\n",
      "Loss: 0.01030:  82%|████████▏ | 206/250 [07:37<01:37,  2.21s/it]\n",
      "Loss: 0.01154:  82%|████████▏ | 206/250 [07:37<01:37,  2.21s/it]\n",
      "Loss: 0.01233:  82%|████████▏ | 206/250 [07:38<01:37,  2.21s/it]\n",
      "Loss: 0.00969:  82%|████████▏ | 206/250 [07:38<01:37,  2.21s/it]\n",
      "Loss: 0.00990:  82%|████████▏ | 206/250 [07:38<01:37,  2.21s/it]\n",
      "Loss: 0.01120:  82%|████████▏ | 206/250 [07:38<01:37,  2.21s/it]\n",
      "Loss: 0.01083:  82%|████████▏ | 206/250 [07:38<01:37,  2.21s/it]\n",
      "Loss: 0.01325:  82%|████████▏ | 206/250 [07:38<01:37,  2.21s/it]\n",
      "Loss: 0.01064:  82%|████████▏ | 206/250 [07:38<01:37,  2.21s/it]\n",
      "Loss: 0.01080:  82%|████████▏ | 206/250 [07:38<01:37,  2.21s/it]\n",
      "Loss: 0.01080:  83%|████████▎ | 207/250 [07:38<01:35,  2.21s/it]\n",
      "Loss: 0.00993:  83%|████████▎ | 207/250 [07:38<01:35,  2.21s/it]\n",
      "Loss: 0.01086:  83%|████████▎ | 207/250 [07:39<01:35,  2.21s/it]\n",
      "Loss: 0.01234:  83%|████████▎ | 207/250 [07:39<01:35,  2.21s/it]\n",
      "Loss: 0.01183:  83%|████████▎ | 207/250 [07:39<01:35,  2.21s/it]\n",
      "Loss: 0.01154:  83%|████████▎ | 207/250 [07:39<01:35,  2.21s/it]\n",
      "Loss: 0.01077:  83%|████████▎ | 207/250 [07:39<01:35,  2.21s/it]\n",
      "Loss: 0.00934:  83%|████████▎ | 207/250 [07:39<01:35,  2.21s/it]\n",
      "Loss: 0.01095:  83%|████████▎ | 207/250 [07:39<01:35,  2.21s/it]\n",
      "Loss: 0.01014:  83%|████████▎ | 207/250 [07:39<01:35,  2.21s/it]\n",
      "Loss: 0.01182:  83%|████████▎ | 207/250 [07:39<01:35,  2.21s/it]\n",
      "Loss: 0.01102:  83%|████████▎ | 207/250 [07:39<01:35,  2.21s/it]\n",
      "Loss: 0.01129:  83%|████████▎ | 207/250 [07:40<01:35,  2.21s/it]\n",
      "Loss: 0.01179:  83%|████████▎ | 207/250 [07:40<01:35,  2.21s/it]\n",
      "Loss: 0.01109:  83%|████████▎ | 207/250 [07:40<01:35,  2.21s/it]\n",
      "Loss: 0.01083:  83%|████████▎ | 207/250 [07:40<01:35,  2.21s/it]\n",
      "Loss: 0.01045:  83%|████████▎ | 207/250 [07:40<01:35,  2.21s/it]\n",
      "Loss: 0.01121:  83%|████████▎ | 207/250 [07:40<01:35,  2.21s/it]\n",
      "Loss: 0.01023:  83%|████████▎ | 207/250 [07:40<01:35,  2.21s/it]\n",
      "Loss: 0.01085:  83%|████████▎ | 207/250 [07:40<01:35,  2.21s/it]\n",
      "Loss: 0.01120:  83%|████████▎ | 207/250 [07:40<01:35,  2.21s/it]\n",
      "Loss: 0.01093:  83%|████████▎ | 207/250 [07:40<01:35,  2.21s/it]\n",
      "Loss: 0.01093:  83%|████████▎ | 208/250 [07:41<01:32,  2.21s/it]\n",
      "Loss: 0.00998:  83%|████████▎ | 208/250 [07:41<01:32,  2.21s/it]\n",
      "Loss: 0.01082:  83%|████████▎ | 208/250 [07:41<01:32,  2.21s/it]\n",
      "Loss: 0.01252:  83%|████████▎ | 208/250 [07:41<01:32,  2.21s/it]\n",
      "Loss: 0.00928:  83%|████████▎ | 208/250 [07:41<01:32,  2.21s/it]\n",
      "Loss: 0.00925:  83%|████████▎ | 208/250 [07:41<01:32,  2.21s/it]\n",
      "Loss: 0.01100:  83%|████████▎ | 208/250 [07:41<01:32,  2.21s/it]\n",
      "Loss: 0.00872:  83%|████████▎ | 208/250 [07:41<01:32,  2.21s/it]\n",
      "Loss: 0.01387:  83%|████████▎ | 208/250 [07:41<01:32,  2.21s/it]\n",
      "Loss: 0.00938:  83%|████████▎ | 208/250 [07:41<01:32,  2.21s/it]\n",
      "Loss: 0.01182:  83%|████████▎ | 208/250 [07:42<01:32,  2.21s/it]\n",
      "Loss: 0.01194:  83%|████████▎ | 208/250 [07:42<01:32,  2.21s/it]\n",
      "Loss: 0.01019:  83%|████████▎ | 208/250 [07:42<01:32,  2.21s/it]\n",
      "Loss: 0.01149:  83%|████████▎ | 208/250 [07:42<01:32,  2.21s/it]\n",
      "Loss: 0.01219:  83%|████████▎ | 208/250 [07:42<01:32,  2.21s/it]\n",
      "Loss: 0.00996:  83%|████████▎ | 208/250 [07:42<01:32,  2.21s/it]\n",
      "Loss: 0.00964:  83%|████████▎ | 208/250 [07:42<01:32,  2.21s/it]\n",
      "Loss: 0.01066:  83%|████████▎ | 208/250 [07:42<01:32,  2.21s/it]\n",
      "Loss: 0.01215:  83%|████████▎ | 208/250 [07:42<01:32,  2.21s/it]\n",
      "Loss: 0.01187:  83%|████████▎ | 208/250 [07:43<01:32,  2.21s/it]\n",
      "Loss: 0.01037:  83%|████████▎ | 208/250 [07:43<01:32,  2.21s/it]\n",
      "Loss: 0.01077:  83%|████████▎ | 208/250 [07:43<01:32,  2.21s/it]\n",
      "Loss: 0.01077:  84%|████████▎ | 209/250 [07:43<01:30,  2.21s/it]\n",
      "Loss: 0.01018:  84%|████████▎ | 209/250 [07:43<01:30,  2.21s/it]\n",
      "Loss: 0.01069:  84%|████████▎ | 209/250 [07:43<01:30,  2.21s/it]\n",
      "Loss: 0.01233:  84%|████████▎ | 209/250 [07:43<01:30,  2.21s/it]\n",
      "Loss: 0.00929:  84%|████████▎ | 209/250 [07:43<01:30,  2.21s/it]\n",
      "Loss: 0.00917:  84%|████████▎ | 209/250 [07:43<01:30,  2.21s/it]\n",
      "Loss: 0.01137:  84%|████████▎ | 209/250 [07:43<01:30,  2.21s/it]\n",
      "Loss: 0.00916:  84%|████████▎ | 209/250 [07:43<01:30,  2.21s/it]\n",
      "Loss: 0.01401:  84%|████████▎ | 209/250 [07:44<01:30,  2.21s/it]\n",
      "Loss: 0.00937:  84%|████████▎ | 209/250 [07:44<01:30,  2.21s/it]\n",
      "Loss: 0.01215:  84%|████████▎ | 209/250 [07:44<01:30,  2.21s/it]\n",
      "Loss: 0.01227:  84%|████████▎ | 209/250 [07:44<01:30,  2.21s/it]\n",
      "Loss: 0.01035:  84%|████████▎ | 209/250 [07:44<01:30,  2.21s/it]\n",
      "Loss: 0.01103:  84%|████████▎ | 209/250 [07:44<01:30,  2.21s/it]\n",
      "Loss: 0.01199:  84%|████████▎ | 209/250 [07:44<01:30,  2.21s/it]\n",
      "Loss: 0.01043:  84%|████████▎ | 209/250 [07:44<01:30,  2.21s/it]\n",
      "Loss: 0.00969:  84%|████████▎ | 209/250 [07:44<01:30,  2.21s/it]\n",
      "Loss: 0.01080:  84%|████████▎ | 209/250 [07:45<01:30,  2.21s/it]\n",
      "Loss: 0.01069:  84%|████████▎ | 209/250 [07:45<01:30,  2.21s/it]\n",
      "Loss: 0.01341:  84%|████████▎ | 209/250 [07:45<01:30,  2.21s/it]\n",
      "Loss: 0.01057:  84%|████████▎ | 209/250 [07:45<01:30,  2.21s/it]\n",
      "Loss: 0.01090:  84%|████████▎ | 209/250 [07:45<01:30,  2.21s/it]\n",
      "Loss: 0.01090:  84%|████████▍ | 210/250 [07:45<01:28,  2.21s/it]\n",
      "Loss: 0.00975:  84%|████████▍ | 210/250 [07:45<01:28,  2.21s/it]\n",
      "Loss: 0.01044:  84%|████████▍ | 210/250 [07:45<01:28,  2.21s/it]\n",
      "Loss: 0.01042:  84%|████████▍ | 210/250 [07:45<01:28,  2.21s/it]\n",
      "Loss: 0.01172:  84%|████████▍ | 210/250 [07:45<01:28,  2.21s/it]\n",
      "Loss: 0.01165:  84%|████████▍ | 210/250 [07:45<01:28,  2.21s/it]\n",
      "Loss: 0.01074:  84%|████████▍ | 210/250 [07:46<01:28,  2.21s/it]\n",
      "Loss: 0.01132:  84%|████████▍ | 210/250 [07:46<01:28,  2.21s/it]\n",
      "Loss: 0.01044:  84%|████████▍ | 210/250 [07:46<01:28,  2.21s/it]\n",
      "Loss: 0.01005:  84%|████████▍ | 210/250 [07:46<01:28,  2.21s/it]\n",
      "Loss: 0.01153:  84%|████████▍ | 210/250 [07:46<01:28,  2.21s/it]\n",
      "Loss: 0.01100:  84%|████████▍ | 210/250 [07:46<01:28,  2.21s/it]\n",
      "Loss: 0.01143:  84%|████████▍ | 210/250 [07:46<01:28,  2.21s/it]\n",
      "Loss: 0.01157:  84%|████████▍ | 210/250 [07:46<01:28,  2.21s/it]\n",
      "Loss: 0.01077:  84%|████████▍ | 210/250 [07:46<01:28,  2.21s/it]\n",
      "Loss: 0.01082:  84%|████████▍ | 210/250 [07:47<01:28,  2.21s/it]\n",
      "Loss: 0.01039:  84%|████████▍ | 210/250 [07:47<01:28,  2.21s/it]\n",
      "Loss: 0.01135:  84%|████████▍ | 210/250 [07:47<01:28,  2.21s/it]\n",
      "Loss: 0.01071:  84%|████████▍ | 210/250 [07:47<01:28,  2.21s/it]\n",
      "Loss: 0.00937:  84%|████████▍ | 210/250 [07:47<01:28,  2.21s/it]\n",
      "Loss: 0.01047:  84%|████████▍ | 210/250 [07:47<01:28,  2.21s/it]\n",
      "Loss: 0.01063:  84%|████████▍ | 210/250 [07:47<01:28,  2.21s/it]\n",
      "Loss: 0.01063:  84%|████████▍ | 211/250 [07:47<01:26,  2.21s/it]\n",
      "Loss: 0.00968:  84%|████████▍ | 211/250 [07:47<01:26,  2.21s/it]\n",
      "Loss: 0.01038:  84%|████████▍ | 211/250 [07:47<01:26,  2.21s/it]\n",
      "Loss: 0.01221:  84%|████████▍ | 211/250 [07:47<01:26,  2.21s/it]\n",
      "Loss: 0.00935:  84%|████████▍ | 211/250 [07:48<01:26,  2.21s/it]\n",
      "Loss: 0.01028:  84%|████████▍ | 211/250 [07:48<01:26,  2.21s/it]\n",
      "Loss: 0.00962:  84%|████████▍ | 211/250 [07:48<01:26,  2.21s/it]\n",
      "Loss: 0.00878:  84%|████████▍ | 211/250 [07:48<01:26,  2.21s/it]\n",
      "Loss: 0.01351:  84%|████████▍ | 211/250 [07:48<01:26,  2.21s/it]\n",
      "Loss: 0.01220:  84%|████████▍ | 211/250 [07:48<01:26,  2.21s/it]\n",
      "Loss: 0.01131:  84%|████████▍ | 211/250 [07:48<01:26,  2.21s/it]\n",
      "Loss: 0.01123:  84%|████████▍ | 211/250 [07:48<01:26,  2.21s/it]\n",
      "Loss: 0.01135:  84%|████████▍ | 211/250 [07:48<01:26,  2.21s/it]\n",
      "Loss: 0.01198:  84%|████████▍ | 211/250 [07:49<01:26,  2.21s/it]\n",
      "Loss: 0.01088:  84%|████████▍ | 211/250 [07:49<01:26,  2.21s/it]\n",
      "Loss: 0.01085:  84%|████████▍ | 211/250 [07:49<01:26,  2.21s/it]\n",
      "Loss: 0.01029:  84%|████████▍ | 211/250 [07:49<01:26,  2.21s/it]\n",
      "Loss: 0.01121:  84%|████████▍ | 211/250 [07:49<01:26,  2.21s/it]\n",
      "Loss: 0.01013:  84%|████████▍ | 211/250 [07:49<01:26,  2.21s/it]\n",
      "Loss: 0.01072:  84%|████████▍ | 211/250 [07:49<01:26,  2.21s/it]\n",
      "Loss: 0.01070:  84%|████████▍ | 211/250 [07:49<01:26,  2.21s/it]\n",
      "Loss: 0.01100:  84%|████████▍ | 211/250 [07:49<01:26,  2.21s/it]\n",
      "Loss: 0.01100:  85%|████████▍ | 212/250 [07:49<01:24,  2.21s/it]\n",
      "Loss: 0.00998:  85%|████████▍ | 212/250 [07:49<01:24,  2.21s/it]\n",
      "Loss: 0.01081:  85%|████████▍ | 212/250 [07:50<01:24,  2.21s/it]\n",
      "Loss: 0.01221:  85%|████████▍ | 212/250 [07:50<01:24,  2.21s/it]\n",
      "Loss: 0.00916:  85%|████████▍ | 212/250 [07:50<01:24,  2.21s/it]\n",
      "Loss: 0.00926:  85%|████████▍ | 212/250 [07:50<01:24,  2.21s/it]\n",
      "Loss: 0.01080:  85%|████████▍ | 212/250 [07:50<01:24,  2.21s/it]\n",
      "Loss: 0.00865:  85%|████████▍ | 212/250 [07:50<01:24,  2.21s/it]\n",
      "Loss: 0.01368:  85%|████████▍ | 212/250 [07:50<01:24,  2.21s/it]\n",
      "Loss: 0.00946:  85%|████████▍ | 212/250 [07:50<01:24,  2.21s/it]\n",
      "Loss: 0.01164:  85%|████████▍ | 212/250 [07:50<01:24,  2.21s/it]\n",
      "Loss: 0.01244:  85%|████████▍ | 212/250 [07:51<01:24,  2.21s/it]\n",
      "Loss: 0.01037:  85%|████████▍ | 212/250 [07:51<01:24,  2.21s/it]\n",
      "Loss: 0.01070:  85%|████████▍ | 212/250 [07:51<01:24,  2.21s/it]\n",
      "Loss: 0.00969:  85%|████████▍ | 212/250 [07:51<01:24,  2.21s/it]\n",
      "Loss: 0.00981:  85%|████████▍ | 212/250 [07:51<01:24,  2.21s/it]\n",
      "Loss: 0.01125:  85%|████████▍ | 212/250 [07:51<01:24,  2.21s/it]\n",
      "Loss: 0.00962:  85%|████████▍ | 212/250 [07:51<01:24,  2.21s/it]\n",
      "Loss: 0.01000:  85%|████████▍ | 212/250 [07:51<01:24,  2.21s/it]\n",
      "Loss: 0.01059:  85%|████████▍ | 212/250 [07:52<01:24,  2.21s/it]\n",
      "Loss: 0.01042:  85%|████████▍ | 212/250 [07:52<01:24,  2.21s/it]\n",
      "Loss: 0.01101:  85%|████████▍ | 212/250 [07:52<01:24,  2.21s/it]\n",
      "Loss: 0.01101:  85%|████████▌ | 213/250 [07:52<01:22,  2.24s/it]\n",
      "Loss: 0.01016:  85%|████████▌ | 213/250 [07:52<01:22,  2.24s/it]\n",
      "Loss: 0.01084:  85%|████████▌ | 213/250 [07:52<01:22,  2.24s/it]\n",
      "Loss: 0.01260:  85%|████████▌ | 213/250 [07:52<01:22,  2.24s/it]\n",
      "Loss: 0.01184:  85%|████████▌ | 213/250 [07:52<01:22,  2.24s/it]\n",
      "Loss: 0.00847:  85%|████████▌ | 213/250 [07:52<01:22,  2.24s/it]\n",
      "Loss: 0.00946:  85%|████████▌ | 213/250 [07:52<01:22,  2.24s/it]\n",
      "Loss: 0.00929:  85%|████████▌ | 213/250 [07:52<01:22,  2.24s/it]\n",
      "Loss: 0.01124:  85%|████████▌ | 213/250 [07:53<01:22,  2.24s/it]\n",
      "Loss: 0.00963:  85%|████████▌ | 213/250 [07:53<01:22,  2.24s/it]\n",
      "Loss: 0.01005:  85%|████████▌ | 213/250 [07:53<01:22,  2.24s/it]\n",
      "Loss: 0.01102:  85%|████████▌ | 213/250 [07:53<01:22,  2.24s/it]\n",
      "Loss: 0.01058:  85%|████████▌ | 213/250 [07:53<01:22,  2.24s/it]\n",
      "Loss: 0.01063:  85%|████████▌ | 213/250 [07:53<01:22,  2.24s/it]\n",
      "Loss: 0.00993:  85%|████████▌ | 213/250 [07:53<01:22,  2.24s/it]\n",
      "Loss: 0.00946:  85%|████████▌ | 213/250 [07:53<01:22,  2.24s/it]\n",
      "Loss: 0.01057:  85%|████████▌ | 213/250 [07:53<01:22,  2.24s/it]\n",
      "Loss: 0.01176:  85%|████████▌ | 213/250 [07:54<01:22,  2.24s/it]\n",
      "Loss: 0.01173:  85%|████████▌ | 213/250 [07:54<01:22,  2.24s/it]\n",
      "Loss: 0.01042:  85%|████████▌ | 213/250 [07:54<01:22,  2.24s/it]\n",
      "Loss: 0.01088:  85%|████████▌ | 213/250 [07:54<01:22,  2.24s/it]\n",
      "Loss: 0.01088:  86%|████████▌ | 214/250 [07:54<01:20,  2.25s/it]\n",
      "Loss: 0.01016:  86%|████████▌ | 214/250 [07:54<01:20,  2.25s/it]\n",
      "Loss: 0.01079:  86%|████████▌ | 214/250 [07:54<01:20,  2.25s/it]\n",
      "Loss: 0.01234:  86%|████████▌ | 214/250 [07:54<01:20,  2.25s/it]\n",
      "Loss: 0.00936:  86%|████████▌ | 214/250 [07:54<01:20,  2.25s/it]\n",
      "Loss: 0.00932:  86%|████████▌ | 214/250 [07:54<01:20,  2.25s/it]\n",
      "Loss: 0.00982:  86%|████████▌ | 214/250 [07:55<01:20,  2.25s/it]\n",
      "Loss: 0.01033:  86%|████████▌ | 214/250 [07:55<01:20,  2.25s/it]\n",
      "Loss: 0.01081:  86%|████████▌ | 214/250 [07:55<01:20,  2.25s/it]\n",
      "Loss: 0.01123:  86%|████████▌ | 214/250 [07:55<01:20,  2.25s/it]\n",
      "Loss: 0.01023:  86%|████████▌ | 214/250 [07:55<01:20,  2.25s/it]\n",
      "Loss: 0.01170:  86%|████████▌ | 214/250 [07:55<01:20,  2.25s/it]\n",
      "Loss: 0.00996:  86%|████████▌ | 214/250 [07:55<01:20,  2.25s/it]\n",
      "Loss: 0.01140:  86%|████████▌ | 214/250 [07:55<01:20,  2.25s/it]\n",
      "Loss: 0.01189:  86%|████████▌ | 214/250 [07:55<01:20,  2.25s/it]\n",
      "Loss: 0.00990:  86%|████████▌ | 214/250 [07:56<01:20,  2.25s/it]\n",
      "Loss: 0.00960:  86%|████████▌ | 214/250 [07:56<01:20,  2.25s/it]\n",
      "Loss: 0.00990:  86%|████████▌ | 214/250 [07:56<01:20,  2.25s/it]\n",
      "Loss: 0.01173:  86%|████████▌ | 214/250 [07:56<01:20,  2.25s/it]\n",
      "Loss: 0.01060:  86%|████████▌ | 214/250 [07:56<01:20,  2.25s/it]\n",
      "Loss: 0.01027:  86%|████████▌ | 214/250 [07:56<01:20,  2.25s/it]\n",
      "Loss: 0.01081:  86%|████████▌ | 214/250 [07:56<01:20,  2.25s/it]\n",
      "Loss: 0.01081:  86%|████████▌ | 215/250 [07:56<01:18,  2.24s/it]\n",
      "Loss: 0.01003:  86%|████████▌ | 215/250 [07:56<01:18,  2.24s/it]\n",
      "Loss: 0.01101:  86%|████████▌ | 215/250 [07:56<01:18,  2.24s/it]\n",
      "Loss: 0.00931:  86%|████████▌ | 215/250 [07:57<01:18,  2.24s/it]\n",
      "Loss: 0.01016:  86%|████████▌ | 215/250 [07:57<01:18,  2.24s/it]\n",
      "Loss: 0.00933:  86%|████████▌ | 215/250 [07:57<01:18,  2.24s/it]\n",
      "Loss: 0.01083:  86%|████████▌ | 215/250 [07:57<01:18,  2.24s/it]\n",
      "Loss: 0.01144:  86%|████████▌ | 215/250 [07:57<01:18,  2.24s/it]\n",
      "Loss: 0.01032:  86%|████████▌ | 215/250 [07:57<01:18,  2.24s/it]\n",
      "Loss: 0.01065:  86%|████████▌ | 215/250 [07:57<01:18,  2.24s/it]\n",
      "Loss: 0.01171:  86%|████████▌ | 215/250 [07:57<01:18,  2.24s/it]\n",
      "Loss: 0.01111:  86%|████████▌ | 215/250 [07:57<01:18,  2.24s/it]\n",
      "Loss: 0.01026:  86%|████████▌ | 215/250 [07:58<01:18,  2.24s/it]\n",
      "Loss: 0.01017:  86%|████████▌ | 215/250 [07:58<01:18,  2.24s/it]\n",
      "Loss: 0.01127:  86%|████████▌ | 215/250 [07:58<01:18,  2.24s/it]\n",
      "Loss: 0.01053:  86%|████████▌ | 215/250 [07:58<01:18,  2.24s/it]\n",
      "Loss: 0.00997:  86%|████████▌ | 215/250 [07:58<01:18,  2.24s/it]\n",
      "Loss: 0.01097:  86%|████████▌ | 215/250 [07:58<01:18,  2.24s/it]\n",
      "Loss: 0.01129:  86%|████████▌ | 215/250 [07:58<01:18,  2.24s/it]\n",
      "Loss: 0.00943:  86%|████████▌ | 215/250 [07:58<01:18,  2.24s/it]\n",
      "Loss: 0.01092:  86%|████████▌ | 215/250 [07:58<01:18,  2.24s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 52.82it/s]\u001b[A\n",
      "Loss: 0.01092:  86%|████████▋ | 216/250 [07:58<01:16,  2.26s/it]\n",
      "Loss: 0.01066:  86%|████████▋ | 216/250 [07:59<01:16,  2.26s/it]\n",
      "Loss: 0.01060:  86%|████████▋ | 216/250 [07:59<01:16,  2.26s/it]\n",
      "Loss: 0.01042:  86%|████████▋ | 216/250 [07:59<01:16,  2.26s/it]\n",
      "Loss: 0.01025:  86%|████████▋ | 216/250 [07:59<01:16,  2.26s/it]\n",
      "Loss: 0.01077:  86%|████████▋ | 216/250 [07:59<01:16,  2.26s/it]\n",
      "Loss: 0.00980:  86%|████████▋ | 216/250 [07:59<01:16,  2.26s/it]\n",
      "Loss: 0.00960:  86%|████████▋ | 216/250 [07:59<01:16,  2.26s/it]\n",
      "Loss: 0.01113:  86%|████████▋ | 216/250 [07:59<01:16,  2.26s/it]\n",
      "Loss: 0.00945:  86%|████████▋ | 216/250 [07:59<01:16,  2.26s/it]\n",
      "Loss: 0.01180:  86%|████████▋ | 216/250 [08:00<01:16,  2.26s/it]\n",
      "Loss: 0.00996:  86%|████████▋ | 216/250 [08:00<01:16,  2.26s/it]\n",
      "Loss: 0.01133:  86%|████████▋ | 216/250 [08:00<01:16,  2.26s/it]\n",
      "Loss: 0.01216:  86%|████████▋ | 216/250 [08:00<01:16,  2.26s/it]\n",
      "Loss: 0.00985:  86%|████████▋ | 216/250 [08:00<01:16,  2.26s/it]\n",
      "Loss: 0.00944:  86%|████████▋ | 216/250 [08:00<01:16,  2.26s/it]\n",
      "Loss: 0.01050:  86%|████████▋ | 216/250 [08:00<01:16,  2.26s/it]\n",
      "Loss: 0.01194:  86%|████████▋ | 216/250 [08:00<01:16,  2.26s/it]\n",
      "Loss: 0.01055:  86%|████████▋ | 216/250 [08:00<01:16,  2.26s/it]\n",
      "Loss: 0.01066:  86%|████████▋ | 216/250 [08:01<01:16,  2.26s/it]\n",
      "Loss: 0.01074:  86%|████████▋ | 216/250 [08:01<01:16,  2.26s/it]\n",
      "Loss: 0.01074:  87%|████████▋ | 217/250 [08:01<01:14,  2.24s/it]\n",
      "Loss: 0.01044:  87%|████████▋ | 217/250 [08:01<01:14,  2.24s/it]\n",
      "Loss: 0.01079:  87%|████████▋ | 217/250 [08:01<01:14,  2.24s/it]\n",
      "Loss: 0.01214:  87%|████████▋ | 217/250 [08:01<01:14,  2.24s/it]\n",
      "Loss: 0.00908:  87%|████████▋ | 217/250 [08:01<01:14,  2.24s/it]\n",
      "Loss: 0.00930:  87%|████████▋ | 217/250 [08:01<01:14,  2.24s/it]\n",
      "Loss: 0.01108:  87%|████████▋ | 217/250 [08:01<01:14,  2.24s/it]\n",
      "Loss: 0.01178:  87%|████████▋ | 217/250 [08:01<01:14,  2.24s/it]\n",
      "Loss: 0.01032:  87%|████████▋ | 217/250 [08:02<01:14,  2.24s/it]\n",
      "Loss: 0.01211:  87%|████████▋ | 217/250 [08:02<01:14,  2.24s/it]\n",
      "Loss: 0.01097:  87%|████████▋ | 217/250 [08:02<01:14,  2.24s/it]\n",
      "Loss: 0.01003:  87%|████████▋ | 217/250 [08:02<01:14,  2.24s/it]\n",
      "Loss: 0.00988:  87%|████████▋ | 217/250 [08:02<01:14,  2.24s/it]\n",
      "Loss: 0.01060:  87%|████████▋ | 217/250 [08:02<01:14,  2.24s/it]\n",
      "Loss: 0.01088:  87%|████████▋ | 217/250 [08:02<01:14,  2.24s/it]\n",
      "Loss: 0.01089:  87%|████████▋ | 217/250 [08:02<01:14,  2.24s/it]\n",
      "Loss: 0.00877:  87%|████████▋ | 217/250 [08:02<01:14,  2.24s/it]\n",
      "Loss: 0.01082:  87%|████████▋ | 217/250 [08:03<01:14,  2.24s/it]\n",
      "Loss: 0.01017:  87%|████████▋ | 217/250 [08:03<01:14,  2.24s/it]\n",
      "Loss: 0.01161:  87%|████████▋ | 217/250 [08:03<01:14,  2.24s/it]\n",
      "Loss: 0.01051:  87%|████████▋ | 217/250 [08:03<01:14,  2.24s/it]\n",
      "Loss: 0.01063:  87%|████████▋ | 217/250 [08:03<01:14,  2.24s/it]\n",
      "Loss: 0.01063:  87%|████████▋ | 218/250 [08:03<01:11,  2.24s/it]\n",
      "Loss: 0.01006:  87%|████████▋ | 218/250 [08:03<01:11,  2.24s/it]\n",
      "Loss: 0.01081:  87%|████████▋ | 218/250 [08:03<01:11,  2.24s/it]\n",
      "Loss: 0.01028:  87%|████████▋ | 218/250 [08:03<01:11,  2.24s/it]\n",
      "Loss: 0.01013:  87%|████████▋ | 218/250 [08:03<01:11,  2.24s/it]\n",
      "Loss: 0.01170:  87%|████████▋ | 218/250 [08:03<01:11,  2.24s/it]\n",
      "Loss: 0.00951:  87%|████████▋ | 218/250 [08:04<01:11,  2.24s/it]\n",
      "Loss: 0.00926:  87%|████████▋ | 218/250 [08:04<01:11,  2.24s/it]\n",
      "Loss: 0.01067:  87%|████████▋ | 218/250 [08:04<01:11,  2.24s/it]\n",
      "Loss: 0.00990:  87%|████████▋ | 218/250 [08:04<01:11,  2.24s/it]\n",
      "Loss: 0.00974:  87%|████████▋ | 218/250 [08:04<01:11,  2.24s/it]\n",
      "Loss: 0.01103:  87%|████████▋ | 218/250 [08:04<01:11,  2.24s/it]\n",
      "Loss: 0.01097:  87%|████████▋ | 218/250 [08:04<01:11,  2.24s/it]\n",
      "Loss: 0.01059:  87%|████████▋ | 218/250 [08:04<01:11,  2.24s/it]\n",
      "Loss: 0.01110:  87%|████████▋ | 218/250 [08:04<01:11,  2.24s/it]\n",
      "Loss: 0.01082:  87%|████████▋ | 218/250 [08:05<01:11,  2.24s/it]\n",
      "Loss: 0.01023:  87%|████████▋ | 218/250 [08:05<01:11,  2.24s/it]\n",
      "Loss: 0.01076:  87%|████████▋ | 218/250 [08:05<01:11,  2.24s/it]\n",
      "Loss: 0.01103:  87%|████████▋ | 218/250 [08:05<01:11,  2.24s/it]\n",
      "Loss: 0.00948:  87%|████████▋ | 218/250 [08:05<01:11,  2.24s/it]\n",
      "Loss: 0.01086:  87%|████████▋ | 218/250 [08:05<01:11,  2.24s/it]\n",
      "Loss: 0.01086:  88%|████████▊ | 219/250 [08:05<01:09,  2.24s/it]\n",
      "Loss: 0.01009:  88%|████████▊ | 219/250 [08:05<01:09,  2.24s/it]\n",
      "Loss: 0.01070:  88%|████████▊ | 219/250 [08:05<01:09,  2.24s/it]\n",
      "Loss: 0.01046:  88%|████████▊ | 219/250 [08:05<01:09,  2.24s/it]\n",
      "Loss: 0.00906:  88%|████████▊ | 219/250 [08:06<01:09,  2.24s/it]\n",
      "Loss: 0.01000:  88%|████████▊ | 219/250 [08:06<01:09,  2.24s/it]\n",
      "Loss: 0.01061:  88%|████████▊ | 219/250 [08:06<01:09,  2.24s/it]\n",
      "Loss: 0.01039:  88%|████████▊ | 219/250 [08:06<01:09,  2.24s/it]\n",
      "Loss: 0.01251:  88%|████████▊ | 219/250 [08:06<01:09,  2.24s/it]\n",
      "Loss: 0.01022:  88%|████████▊ | 219/250 [08:06<01:09,  2.24s/it]\n",
      "Loss: 0.01055:  88%|████████▊ | 219/250 [08:06<01:09,  2.24s/it]\n",
      "Loss: 0.01217:  88%|████████▊ | 219/250 [08:06<01:09,  2.24s/it]\n",
      "Loss: 0.01009:  88%|████████▊ | 219/250 [08:06<01:09,  2.24s/it]\n",
      "Loss: 0.01124:  88%|████████▊ | 219/250 [08:07<01:09,  2.24s/it]\n",
      "Loss: 0.01241:  88%|████████▊ | 219/250 [08:07<01:09,  2.24s/it]\n",
      "Loss: 0.01015:  88%|████████▊ | 219/250 [08:07<01:09,  2.24s/it]\n",
      "Loss: 0.00959:  88%|████████▊ | 219/250 [08:07<01:09,  2.24s/it]\n",
      "Loss: 0.01053:  88%|████████▊ | 219/250 [08:07<01:09,  2.24s/it]\n",
      "Loss: 0.01118:  88%|████████▊ | 219/250 [08:07<01:09,  2.24s/it]\n",
      "Loss: 0.01312:  88%|████████▊ | 219/250 [08:07<01:09,  2.24s/it]\n",
      "Loss: 0.01070:  88%|████████▊ | 219/250 [08:07<01:09,  2.24s/it]\n",
      "Loss: 0.01086:  88%|████████▊ | 219/250 [08:07<01:09,  2.24s/it]\n",
      "Loss: 0.01086:  88%|████████▊ | 220/250 [08:07<01:07,  2.25s/it]\n",
      "Loss: 0.01010:  88%|████████▊ | 220/250 [08:08<01:07,  2.25s/it]\n",
      "Loss: 0.01095:  88%|████████▊ | 220/250 [08:08<01:07,  2.25s/it]\n",
      "Loss: 0.01215:  88%|████████▊ | 220/250 [08:08<01:07,  2.25s/it]\n",
      "Loss: 0.01163:  88%|████████▊ | 220/250 [08:08<01:07,  2.25s/it]\n",
      "Loss: 0.01157:  88%|████████▊ | 220/250 [08:08<01:07,  2.25s/it]\n",
      "Loss: 0.01073:  88%|████████▊ | 220/250 [08:08<01:07,  2.25s/it]\n",
      "Loss: 0.00928:  88%|████████▊ | 220/250 [08:08<01:07,  2.25s/it]\n",
      "Loss: 0.01095:  88%|████████▊ | 220/250 [08:08<01:07,  2.25s/it]\n",
      "Loss: 0.00965:  88%|████████▊ | 220/250 [08:08<01:07,  2.25s/it]\n",
      "Loss: 0.01141:  88%|████████▊ | 220/250 [08:09<01:07,  2.25s/it]\n",
      "Loss: 0.01091:  88%|████████▊ | 220/250 [08:09<01:07,  2.25s/it]\n",
      "Loss: 0.01048:  88%|████████▊ | 220/250 [08:09<01:07,  2.25s/it]\n",
      "Loss: 0.00982:  88%|████████▊ | 220/250 [08:09<01:07,  2.25s/it]\n",
      "Loss: 0.00963:  88%|████████▊ | 220/250 [08:09<01:07,  2.25s/it]\n",
      "Loss: 0.00942:  88%|████████▊ | 220/250 [08:09<01:07,  2.25s/it]\n",
      "Loss: 0.00930:  88%|████████▊ | 220/250 [08:09<01:07,  2.25s/it]\n",
      "Loss: 0.01091:  88%|████████▊ | 220/250 [08:09<01:07,  2.25s/it]\n",
      "Loss: 0.00987:  88%|████████▊ | 220/250 [08:09<01:07,  2.25s/it]\n",
      "Loss: 0.01134:  88%|████████▊ | 220/250 [08:10<01:07,  2.25s/it]\n",
      "Loss: 0.01098:  88%|████████▊ | 220/250 [08:10<01:07,  2.25s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 56.33it/s]\u001b[A\n",
      "Loss: 0.01098:  88%|████████▊ | 221/250 [08:10<01:05,  2.25s/it]\n",
      "Loss: 0.01071:  88%|████████▊ | 221/250 [08:10<01:05,  2.25s/it]\n",
      "Loss: 0.01062:  88%|████████▊ | 221/250 [08:10<01:05,  2.25s/it]\n",
      "Loss: 0.01202:  88%|████████▊ | 221/250 [08:10<01:05,  2.25s/it]\n",
      "Loss: 0.00920:  88%|████████▊ | 221/250 [08:10<01:05,  2.25s/it]\n",
      "Loss: 0.00914:  88%|████████▊ | 221/250 [08:10<01:05,  2.25s/it]\n",
      "Loss: 0.01093:  88%|████████▊ | 221/250 [08:10<01:05,  2.25s/it]\n",
      "Loss: 0.00845:  88%|████████▊ | 221/250 [08:10<01:05,  2.25s/it]\n",
      "Loss: 0.01003:  88%|████████▊ | 221/250 [08:11<01:05,  2.25s/it]\n",
      "Loss: 0.00987:  88%|████████▊ | 221/250 [08:11<01:05,  2.25s/it]\n",
      "Loss: 0.01155:  88%|████████▊ | 221/250 [08:11<01:05,  2.25s/it]\n",
      "Loss: 0.01116:  88%|████████▊ | 221/250 [08:11<01:05,  2.25s/it]\n",
      "Loss: 0.00975:  88%|████████▊ | 221/250 [08:11<01:05,  2.25s/it]\n",
      "Loss: 0.01004:  88%|████████▊ | 221/250 [08:11<01:05,  2.25s/it]\n",
      "Loss: 0.00970:  88%|████████▊ | 221/250 [08:11<01:05,  2.25s/it]\n",
      "Loss: 0.01121:  88%|████████▊ | 221/250 [08:11<01:05,  2.25s/it]\n",
      "Loss: 0.00961:  88%|████████▊ | 221/250 [08:11<01:05,  2.25s/it]\n",
      "Loss: 0.01100:  88%|████████▊ | 221/250 [08:12<01:05,  2.25s/it]\n",
      "Loss: 0.01081:  88%|████████▊ | 221/250 [08:12<01:05,  2.25s/it]\n",
      "Loss: 0.01055:  88%|████████▊ | 221/250 [08:12<01:05,  2.25s/it]\n",
      "Loss: 0.01067:  88%|████████▊ | 221/250 [08:12<01:05,  2.25s/it]\n",
      "Loss: 0.01067:  89%|████████▉ | 222/250 [08:12<01:03,  2.26s/it]\n",
      "Loss: 0.01022:  89%|████████▉ | 222/250 [08:12<01:03,  2.26s/it]\n",
      "Loss: 0.01046:  89%|████████▉ | 222/250 [08:12<01:03,  2.26s/it]\n",
      "Loss: 0.01158:  89%|████████▉ | 222/250 [08:12<01:03,  2.26s/it]\n",
      "Loss: 0.00966:  89%|████████▉ | 222/250 [08:12<01:03,  2.26s/it]\n",
      "Loss: 0.00983:  89%|████████▉ | 222/250 [08:12<01:03,  2.26s/it]\n",
      "Loss: 0.00977:  89%|████████▉ | 222/250 [08:13<01:03,  2.26s/it]\n",
      "Loss: 0.01069:  89%|████████▉ | 222/250 [08:13<01:03,  2.26s/it]\n",
      "Loss: 0.01364:  89%|████████▉ | 222/250 [08:13<01:03,  2.26s/it]\n",
      "Loss: 0.00918:  89%|████████▉ | 222/250 [08:13<01:03,  2.26s/it]\n",
      "Loss: 0.01171:  89%|████████▉ | 222/250 [08:13<01:03,  2.26s/it]\n",
      "Loss: 0.01213:  89%|████████▉ | 222/250 [08:13<01:03,  2.26s/it]\n",
      "Loss: 0.00980:  89%|████████▉ | 222/250 [08:13<01:03,  2.26s/it]\n",
      "Loss: 0.01162:  89%|████████▉ | 222/250 [08:13<01:03,  2.26s/it]\n",
      "Loss: 0.00968:  89%|████████▉ | 222/250 [08:13<01:03,  2.26s/it]\n",
      "Loss: 0.00926:  89%|████████▉ | 222/250 [08:14<01:03,  2.26s/it]\n",
      "Loss: 0.01025:  89%|████████▉ | 222/250 [08:14<01:03,  2.26s/it]\n",
      "Loss: 0.01087:  89%|████████▉ | 222/250 [08:14<01:03,  2.26s/it]\n",
      "Loss: 0.01098:  89%|████████▉ | 222/250 [08:14<01:03,  2.26s/it]\n",
      "Loss: 0.00932:  89%|████████▉ | 222/250 [08:14<01:03,  2.26s/it]\n",
      "Loss: 0.01109:  89%|████████▉ | 222/250 [08:14<01:03,  2.26s/it]\n",
      "Loss: 0.01109:  89%|████████▉ | 223/250 [08:14<01:00,  2.24s/it]\n",
      "Loss: 0.01006:  89%|████████▉ | 223/250 [08:14<01:00,  2.24s/it]\n",
      "Loss: 0.01096:  89%|████████▉ | 223/250 [08:14<01:00,  2.24s/it]\n",
      "Loss: 0.01231:  89%|████████▉ | 223/250 [08:14<01:00,  2.24s/it]\n",
      "Loss: 0.01184:  89%|████████▉ | 223/250 [08:15<01:00,  2.24s/it]\n",
      "Loss: 0.01207:  89%|████████▉ | 223/250 [08:15<01:00,  2.24s/it]\n",
      "Loss: 0.01096:  89%|████████▉ | 223/250 [08:15<01:00,  2.24s/it]\n",
      "Loss: 0.00861:  89%|████████▉ | 223/250 [08:15<01:00,  2.24s/it]\n",
      "Loss: 0.01038:  89%|████████▉ | 223/250 [08:15<01:00,  2.24s/it]\n",
      "Loss: 0.01008:  89%|████████▉ | 223/250 [08:15<01:00,  2.24s/it]\n",
      "Loss: 0.01121:  89%|████████▉ | 223/250 [08:15<01:00,  2.24s/it]\n",
      "Loss: 0.01112:  89%|████████▉ | 223/250 [08:15<01:00,  2.24s/it]\n",
      "Loss: 0.01021:  89%|████████▉ | 223/250 [08:15<01:00,  2.24s/it]\n",
      "Loss: 0.01022:  89%|████████▉ | 223/250 [08:16<01:00,  2.24s/it]\n",
      "Loss: 0.00957:  89%|████████▉ | 223/250 [08:16<01:00,  2.24s/it]\n",
      "Loss: 0.01120:  89%|████████▉ | 223/250 [08:16<01:00,  2.24s/it]\n",
      "Loss: 0.00955:  89%|████████▉ | 223/250 [08:16<01:00,  2.24s/it]\n",
      "Loss: 0.01078:  89%|████████▉ | 223/250 [08:16<01:00,  2.24s/it]\n",
      "Loss: 0.01018:  89%|████████▉ | 223/250 [08:16<01:00,  2.24s/it]\n",
      "Loss: 0.01160:  89%|████████▉ | 223/250 [08:16<01:00,  2.24s/it]\n",
      "Loss: 0.01040:  89%|████████▉ | 223/250 [08:16<01:00,  2.24s/it]\n",
      "Loss: 0.01084:  89%|████████▉ | 223/250 [08:16<01:00,  2.24s/it]\n",
      "Loss: 0.01084:  90%|████████▉ | 224/250 [08:16<00:58,  2.25s/it]\n",
      "Loss: 0.00993:  90%|████████▉ | 224/250 [08:17<00:58,  2.25s/it]\n",
      "Loss: 0.01085:  90%|████████▉ | 224/250 [08:17<00:58,  2.25s/it]\n",
      "Loss: 0.01171:  90%|████████▉ | 224/250 [08:17<00:58,  2.25s/it]\n",
      "Loss: 0.00995:  90%|████████▉ | 224/250 [08:17<00:58,  2.25s/it]\n",
      "Loss: 0.00922:  90%|████████▉ | 224/250 [08:17<00:58,  2.25s/it]\n",
      "Loss: 0.01085:  90%|████████▉ | 224/250 [08:17<00:58,  2.25s/it]\n",
      "Loss: 0.00865:  90%|████████▉ | 224/250 [08:17<00:58,  2.25s/it]\n",
      "Loss: 0.01007:  90%|████████▉ | 224/250 [08:17<00:58,  2.25s/it]\n",
      "Loss: 0.01209:  90%|████████▉ | 224/250 [08:17<00:58,  2.25s/it]\n",
      "Loss: 0.01096:  90%|████████▉ | 224/250 [08:18<00:58,  2.25s/it]\n",
      "Loss: 0.00978:  90%|████████▉ | 224/250 [08:18<00:58,  2.25s/it]\n",
      "Loss: 0.00976:  90%|████████▉ | 224/250 [08:18<00:58,  2.25s/it]\n",
      "Loss: 0.01200:  90%|████████▉ | 224/250 [08:18<00:58,  2.25s/it]\n",
      "Loss: 0.01071:  90%|████████▉ | 224/250 [08:18<00:58,  2.25s/it]\n",
      "Loss: 0.01064:  90%|████████▉ | 224/250 [08:18<00:58,  2.25s/it]\n",
      "Loss: 0.01015:  90%|████████▉ | 224/250 [08:18<00:58,  2.25s/it]\n",
      "Loss: 0.01091:  90%|████████▉ | 224/250 [08:18<00:58,  2.25s/it]\n",
      "Loss: 0.01050:  90%|████████▉ | 224/250 [08:18<00:58,  2.25s/it]\n",
      "Loss: 0.00943:  90%|████████▉ | 224/250 [08:19<00:58,  2.25s/it]\n",
      "Loss: 0.01015:  90%|████████▉ | 224/250 [08:19<00:58,  2.25s/it]\n",
      "Loss: 0.01061:  90%|████████▉ | 224/250 [08:19<00:58,  2.25s/it]\n",
      "Loss: 0.01061:  90%|█████████ | 225/250 [08:19<00:56,  2.26s/it]\n",
      "Loss: 0.00997:  90%|█████████ | 225/250 [08:19<00:56,  2.26s/it]\n",
      "Loss: 0.01037:  90%|█████████ | 225/250 [08:19<00:56,  2.26s/it]\n",
      "Loss: 0.01023:  90%|█████████ | 225/250 [08:19<00:56,  2.26s/it]\n",
      "Loss: 0.01013:  90%|█████████ | 225/250 [08:19<00:56,  2.26s/it]\n",
      "Loss: 0.00870:  90%|█████████ | 225/250 [08:19<00:56,  2.26s/it]\n",
      "Loss: 0.01082:  90%|█████████ | 225/250 [08:19<00:56,  2.26s/it]\n",
      "Loss: 0.00849:  90%|█████████ | 225/250 [08:19<00:56,  2.26s/it]\n",
      "Loss: 0.01373:  90%|█████████ | 225/250 [08:20<00:56,  2.26s/it]\n",
      "Loss: 0.01217:  90%|█████████ | 225/250 [08:20<00:56,  2.26s/it]\n",
      "Loss: 0.01122:  90%|█████████ | 225/250 [08:20<00:56,  2.26s/it]\n",
      "Loss: 0.00976:  90%|█████████ | 225/250 [08:20<00:56,  2.26s/it]\n",
      "Loss: 0.00996:  90%|█████████ | 225/250 [08:20<00:56,  2.26s/it]\n",
      "Loss: 0.01167:  90%|█████████ | 225/250 [08:20<00:56,  2.26s/it]\n",
      "Loss: 0.01072:  90%|█████████ | 225/250 [08:20<00:56,  2.26s/it]\n",
      "Loss: 0.01033:  90%|█████████ | 225/250 [08:20<00:56,  2.26s/it]\n",
      "Loss: 0.01034:  90%|█████████ | 225/250 [08:20<00:56,  2.26s/it]\n",
      "Loss: 0.01088:  90%|█████████ | 225/250 [08:21<00:56,  2.26s/it]\n",
      "Loss: 0.01073:  90%|█████████ | 225/250 [08:21<00:56,  2.26s/it]\n",
      "Loss: 0.00957:  90%|█████████ | 225/250 [08:21<00:56,  2.26s/it]\n",
      "Loss: 0.01007:  90%|█████████ | 225/250 [08:21<00:56,  2.26s/it]\n",
      "Loss: 0.01074:  90%|█████████ | 225/250 [08:21<00:56,  2.26s/it]\n",
      "Loss: 0.01074:  90%|█████████ | 226/250 [08:21<00:54,  2.25s/it]\n",
      "Loss: 0.00981:  90%|█████████ | 226/250 [08:21<00:54,  2.25s/it]\n",
      "Loss: 0.01067:  90%|█████████ | 226/250 [08:21<00:54,  2.25s/it]\n",
      "Loss: 0.01194:  90%|█████████ | 226/250 [08:21<00:54,  2.25s/it]\n",
      "Loss: 0.00921:  90%|█████████ | 226/250 [08:21<00:54,  2.25s/it]\n",
      "Loss: 0.01174:  90%|█████████ | 226/250 [08:22<00:54,  2.25s/it]\n",
      "Loss: 0.00942:  90%|█████████ | 226/250 [08:22<00:54,  2.25s/it]\n",
      "Loss: 0.00963:  90%|█████████ | 226/250 [08:22<00:54,  2.25s/it]\n",
      "Loss: 0.01131:  90%|█████████ | 226/250 [08:22<00:54,  2.25s/it]\n",
      "Loss: 0.00961:  90%|█████████ | 226/250 [08:22<00:54,  2.25s/it]\n",
      "Loss: 0.00953:  90%|█████████ | 226/250 [08:22<00:54,  2.25s/it]\n",
      "Loss: 0.01101:  90%|█████████ | 226/250 [08:22<00:54,  2.25s/it]\n",
      "Loss: 0.01009:  90%|█████████ | 226/250 [08:22<00:54,  2.25s/it]\n",
      "Loss: 0.01010:  90%|█████████ | 226/250 [08:22<00:54,  2.25s/it]\n",
      "Loss: 0.01100:  90%|█████████ | 226/250 [08:23<00:54,  2.25s/it]\n",
      "Loss: 0.01082:  90%|█████████ | 226/250 [08:23<00:54,  2.25s/it]\n",
      "Loss: 0.01047:  90%|█████████ | 226/250 [08:23<00:54,  2.25s/it]\n",
      "Loss: 0.01059:  90%|█████████ | 226/250 [08:23<00:54,  2.25s/it]\n",
      "Loss: 0.01283:  90%|█████████ | 226/250 [08:23<00:54,  2.25s/it]\n",
      "Loss: 0.01051:  90%|█████████ | 226/250 [08:23<00:54,  2.25s/it]\n",
      "Loss: 0.01056:  90%|█████████ | 226/250 [08:23<00:54,  2.25s/it]\n",
      "Loss: 0.01056:  91%|█████████ | 227/250 [08:23<00:51,  2.26s/it]\n",
      "Loss: 0.01065:  91%|█████████ | 227/250 [08:23<00:51,  2.26s/it]\n",
      "Loss: 0.00947:  91%|█████████ | 227/250 [08:23<00:51,  2.26s/it]\n",
      "Loss: 0.01199:  91%|█████████ | 227/250 [08:24<00:51,  2.26s/it]\n",
      "Loss: 0.00989:  91%|█████████ | 227/250 [08:24<00:51,  2.26s/it]\n",
      "Loss: 0.01020:  91%|█████████ | 227/250 [08:24<00:51,  2.26s/it]\n",
      "Loss: 0.01024:  91%|█████████ | 227/250 [08:24<00:51,  2.26s/it]\n",
      "Loss: 0.00993:  91%|█████████ | 227/250 [08:24<00:51,  2.26s/it]\n",
      "Loss: 0.01114:  91%|█████████ | 227/250 [08:24<00:51,  2.26s/it]\n",
      "Loss: 0.01046:  91%|█████████ | 227/250 [08:24<00:51,  2.26s/it]\n",
      "Loss: 0.01175:  91%|█████████ | 227/250 [08:24<00:51,  2.26s/it]\n",
      "Loss: 0.00973:  91%|█████████ | 227/250 [08:24<00:51,  2.26s/it]\n",
      "Loss: 0.01026:  91%|█████████ | 227/250 [08:25<00:51,  2.26s/it]\n",
      "Loss: 0.01075:  91%|█████████ | 227/250 [08:25<00:51,  2.26s/it]\n",
      "Loss: 0.01059:  91%|█████████ | 227/250 [08:25<00:51,  2.26s/it]\n",
      "Loss: 0.01019:  91%|█████████ | 227/250 [08:25<00:51,  2.26s/it]\n",
      "Loss: 0.01095:  91%|█████████ | 227/250 [08:25<00:51,  2.26s/it]\n",
      "Loss: 0.01060:  91%|█████████ | 227/250 [08:25<00:51,  2.26s/it]\n",
      "Loss: 0.01148:  91%|█████████ | 227/250 [08:25<00:51,  2.26s/it]\n",
      "Loss: 0.01035:  91%|█████████ | 227/250 [08:25<00:51,  2.26s/it]\n",
      "Loss: 0.01088:  91%|█████████ | 227/250 [08:25<00:51,  2.26s/it]\n",
      "Loss: 0.01088:  91%|█████████ | 228/250 [08:25<00:49,  2.24s/it]\n",
      "Loss: 0.00986:  91%|█████████ | 228/250 [08:25<00:49,  2.24s/it]\n",
      "Loss: 0.01044:  91%|█████████ | 228/250 [08:26<00:49,  2.24s/it]\n",
      "Loss: 0.01025:  91%|█████████ | 228/250 [08:26<00:49,  2.24s/it]\n",
      "Loss: 0.01153:  91%|█████████ | 228/250 [08:26<00:49,  2.24s/it]\n",
      "Loss: 0.00905:  91%|█████████ | 228/250 [08:26<00:49,  2.24s/it]\n",
      "Loss: 0.01055:  91%|█████████ | 228/250 [08:26<00:49,  2.24s/it]\n",
      "Loss: 0.00863:  91%|█████████ | 228/250 [08:26<00:49,  2.24s/it]\n",
      "Loss: 0.01024:  91%|█████████ | 228/250 [08:26<00:49,  2.24s/it]\n",
      "Loss: 0.01211:  91%|█████████ | 228/250 [08:26<00:49,  2.24s/it]\n",
      "Loss: 0.01145:  91%|█████████ | 228/250 [08:26<00:49,  2.24s/it]\n",
      "Loss: 0.01118:  91%|█████████ | 228/250 [08:27<00:49,  2.24s/it]\n",
      "Loss: 0.01102:  91%|█████████ | 228/250 [08:27<00:49,  2.24s/it]\n",
      "Loss: 0.01199:  91%|█████████ | 228/250 [08:27<00:49,  2.24s/it]\n",
      "Loss: 0.01096:  91%|█████████ | 228/250 [08:27<00:49,  2.24s/it]\n",
      "Loss: 0.01083:  91%|█████████ | 228/250 [08:27<00:49,  2.24s/it]\n",
      "Loss: 0.00941:  91%|█████████ | 228/250 [08:27<00:49,  2.24s/it]\n",
      "Loss: 0.01081:  91%|█████████ | 228/250 [08:27<00:49,  2.24s/it]\n",
      "Loss: 0.01073:  91%|█████████ | 228/250 [08:27<00:49,  2.24s/it]\n",
      "Loss: 0.01048:  91%|█████████ | 228/250 [08:28<00:49,  2.24s/it]\n",
      "Loss: 0.01062:  91%|█████████ | 228/250 [08:28<00:49,  2.24s/it]\n",
      "Loss: 0.01062:  92%|█████████▏| 229/250 [08:28<00:46,  2.24s/it]\n",
      "Loss: 0.00958:  92%|█████████▏| 229/250 [08:28<00:46,  2.24s/it]\n",
      "Loss: 0.01058:  92%|█████████▏| 229/250 [08:28<00:46,  2.24s/it]\n",
      "Loss: 0.01199:  92%|█████████▏| 229/250 [08:28<00:46,  2.24s/it]\n",
      "Loss: 0.01150:  92%|█████████▏| 229/250 [08:28<00:46,  2.24s/it]\n",
      "Loss: 0.01163:  92%|█████████▏| 229/250 [08:28<00:46,  2.24s/it]\n",
      "Loss: 0.01075:  92%|█████████▏| 229/250 [08:28<00:46,  2.24s/it]\n",
      "Loss: 0.00910:  92%|█████████▏| 229/250 [08:28<00:46,  2.24s/it]\n",
      "Loss: 0.01098:  92%|█████████▏| 229/250 [08:29<00:46,  2.24s/it]\n",
      "Loss: 0.01020:  92%|█████████▏| 229/250 [08:29<00:46,  2.24s/it]\n",
      "Loss: 0.01155:  92%|█████████▏| 229/250 [08:29<00:46,  2.24s/it]\n",
      "Loss: 0.01083:  92%|█████████▏| 229/250 [08:29<00:46,  2.24s/it]\n",
      "Loss: 0.00997:  92%|█████████▏| 229/250 [08:29<00:46,  2.24s/it]\n",
      "Loss: 0.00998:  92%|█████████▏| 229/250 [08:29<00:46,  2.24s/it]\n",
      "Loss: 0.01087:  92%|█████████▏| 229/250 [08:29<00:46,  2.24s/it]\n",
      "Loss: 0.01049:  92%|█████████▏| 229/250 [08:29<00:46,  2.24s/it]\n",
      "Loss: 0.00990:  92%|█████████▏| 229/250 [08:29<00:46,  2.24s/it]\n",
      "Loss: 0.01139:  92%|█████████▏| 229/250 [08:30<00:46,  2.24s/it]\n",
      "Loss: 0.01070:  92%|█████████▏| 229/250 [08:30<00:46,  2.24s/it]\n",
      "Loss: 0.01055:  92%|█████████▏| 229/250 [08:30<00:46,  2.24s/it]\n",
      "Loss: 0.01057:  92%|█████████▏| 229/250 [08:30<00:46,  2.24s/it]\n",
      "Loss: 0.01057:  92%|█████████▏| 230/250 [08:30<00:44,  2.24s/it]\n",
      "Loss: 0.00968:  92%|█████████▏| 230/250 [08:30<00:44,  2.24s/it]\n",
      "Loss: 0.01064:  92%|█████████▏| 230/250 [08:30<00:44,  2.24s/it]\n",
      "Loss: 0.01019:  92%|█████████▏| 230/250 [08:30<00:44,  2.24s/it]\n",
      "Loss: 0.01196:  92%|█████████▏| 230/250 [08:30<00:44,  2.24s/it]\n",
      "Loss: 0.00925:  92%|█████████▏| 230/250 [08:30<00:44,  2.24s/it]\n",
      "Loss: 0.00961:  92%|█████████▏| 230/250 [08:31<00:44,  2.24s/it]\n",
      "Loss: 0.00846:  92%|█████████▏| 230/250 [08:31<00:44,  2.24s/it]\n",
      "Loss: 0.01343:  92%|█████████▏| 230/250 [08:31<00:44,  2.24s/it]\n",
      "Loss: 0.00914:  92%|█████████▏| 230/250 [08:31<00:44,  2.24s/it]\n",
      "Loss: 0.01200:  92%|█████████▏| 230/250 [08:31<00:44,  2.24s/it]\n",
      "Loss: 0.01203:  92%|█████████▏| 230/250 [08:31<00:44,  2.24s/it]\n",
      "Loss: 0.01019:  92%|█████████▏| 230/250 [08:31<00:44,  2.24s/it]\n",
      "Loss: 0.01125:  92%|█████████▏| 230/250 [08:31<00:44,  2.24s/it]\n",
      "Loss: 0.01172:  92%|█████████▏| 230/250 [08:31<00:44,  2.24s/it]\n",
      "Loss: 0.00954:  92%|█████████▏| 230/250 [08:32<00:44,  2.24s/it]\n",
      "Loss: 0.01008:  92%|█████████▏| 230/250 [08:32<00:44,  2.24s/it]\n",
      "Loss: 0.01093:  92%|█████████▏| 230/250 [08:32<00:44,  2.24s/it]\n",
      "Loss: 0.01057:  92%|█████████▏| 230/250 [08:32<00:44,  2.24s/it]\n",
      "Loss: 0.01156:  92%|█████████▏| 230/250 [08:32<00:44,  2.24s/it]\n",
      "Loss: 0.01032:  92%|█████████▏| 230/250 [08:32<00:44,  2.24s/it]\n",
      "Loss: 0.01069:  92%|█████████▏| 230/250 [08:32<00:44,  2.24s/it]\n",
      "Loss: 0.01069:  92%|█████████▏| 231/250 [08:32<00:42,  2.25s/it]\n",
      "Loss: 0.00984:  92%|█████████▏| 231/250 [08:32<00:42,  2.25s/it]\n",
      "Loss: 0.01044:  92%|█████████▏| 231/250 [08:32<00:42,  2.25s/it]\n",
      "Loss: 0.01050:  92%|█████████▏| 231/250 [08:32<00:42,  2.25s/it]\n",
      "Loss: 0.00985:  92%|█████████▏| 231/250 [08:33<00:42,  2.25s/it]\n",
      "Loss: 0.00846:  92%|█████████▏| 231/250 [08:33<00:42,  2.25s/it]\n",
      "Loss: 0.00957:  92%|█████████▏| 231/250 [08:33<00:42,  2.25s/it]\n",
      "Loss: 0.01062:  92%|█████████▏| 231/250 [08:33<00:42,  2.25s/it]\n",
      "Loss: 0.01129:  92%|█████████▏| 231/250 [08:33<00:42,  2.25s/it]\n",
      "Loss: 0.00930:  92%|█████████▏| 231/250 [08:33<00:42,  2.25s/it]\n",
      "Loss: 0.00995:  92%|█████████▏| 231/250 [08:33<00:42,  2.25s/it]\n",
      "Loss: 0.01108:  92%|█████████▏| 231/250 [08:33<00:42,  2.25s/it]\n",
      "Loss: 0.01107:  92%|█████████▏| 231/250 [08:33<00:42,  2.25s/it]\n",
      "Loss: 0.00990:  92%|█████████▏| 231/250 [08:34<00:42,  2.25s/it]\n",
      "Loss: 0.00918:  92%|█████████▏| 231/250 [08:34<00:42,  2.25s/it]\n",
      "Loss: 0.01075:  92%|█████████▏| 231/250 [08:34<00:42,  2.25s/it]\n",
      "Loss: 0.00992:  92%|█████████▏| 231/250 [08:34<00:42,  2.25s/it]\n",
      "Loss: 0.01089:  92%|█████████▏| 231/250 [08:34<00:42,  2.25s/it]\n",
      "Loss: 0.01074:  92%|█████████▏| 231/250 [08:34<00:42,  2.25s/it]\n",
      "Loss: 0.00958:  92%|█████████▏| 231/250 [08:34<00:42,  2.25s/it]\n",
      "Loss: 0.01059:  92%|█████████▏| 231/250 [08:34<00:42,  2.25s/it]\n",
      "Loss: 0.01059:  93%|█████████▎| 232/250 [08:34<00:40,  2.24s/it]\n",
      "Loss: 0.01057:  93%|█████████▎| 232/250 [08:34<00:40,  2.24s/it]\n",
      "Loss: 0.00942:  93%|█████████▎| 232/250 [08:35<00:40,  2.24s/it]\n",
      "Loss: 0.01015:  93%|█████████▎| 232/250 [08:35<00:40,  2.24s/it]\n",
      "Loss: 0.01009:  93%|█████████▎| 232/250 [08:35<00:40,  2.24s/it]\n",
      "Loss: 0.00827:  93%|█████████▎| 232/250 [08:35<00:40,  2.24s/it]\n",
      "Loss: 0.00924:  93%|█████████▎| 232/250 [08:35<00:40,  2.24s/it]\n",
      "Loss: 0.01140:  93%|█████████▎| 232/250 [08:35<00:40,  2.24s/it]\n",
      "Loss: 0.01030:  93%|█████████▎| 232/250 [08:35<00:40,  2.24s/it]\n",
      "Loss: 0.01209:  93%|█████████▎| 232/250 [08:35<00:40,  2.24s/it]\n",
      "Loss: 0.01090:  93%|█████████▎| 232/250 [08:35<00:40,  2.24s/it]\n",
      "Loss: 0.01016:  93%|█████████▎| 232/250 [08:36<00:40,  2.24s/it]\n",
      "Loss: 0.01110:  93%|█████████▎| 232/250 [08:36<00:40,  2.24s/it]\n",
      "Loss: 0.01126:  93%|█████████▎| 232/250 [08:36<00:40,  2.24s/it]\n",
      "Loss: 0.01095:  93%|█████████▎| 232/250 [08:36<00:40,  2.24s/it]\n",
      "Loss: 0.01057:  93%|█████████▎| 232/250 [08:36<00:40,  2.24s/it]\n",
      "Loss: 0.01003:  93%|█████████▎| 232/250 [08:36<00:40,  2.24s/it]\n",
      "Loss: 0.01108:  93%|█████████▎| 232/250 [08:36<00:40,  2.24s/it]\n",
      "Loss: 0.01064:  93%|█████████▎| 232/250 [08:36<00:40,  2.24s/it]\n",
      "Loss: 0.00967:  93%|█████████▎| 232/250 [08:36<00:40,  2.24s/it]\n",
      "Loss: 0.01032:  93%|█████████▎| 232/250 [08:37<00:40,  2.24s/it]\n",
      "Loss: 0.01099:  93%|█████████▎| 232/250 [08:37<00:40,  2.24s/it]\n",
      "Loss: 0.01099:  93%|█████████▎| 233/250 [08:37<00:37,  2.23s/it]\n",
      "Loss: 0.00987:  93%|█████████▎| 233/250 [08:37<00:37,  2.23s/it]\n",
      "Loss: 0.01078:  93%|█████████▎| 233/250 [08:37<00:37,  2.23s/it]\n",
      "Loss: 0.01176:  93%|█████████▎| 233/250 [08:37<00:37,  2.23s/it]\n",
      "Loss: 0.00955:  93%|█████████▎| 233/250 [08:37<00:37,  2.23s/it]\n",
      "Loss: 0.00905:  93%|█████████▎| 233/250 [08:37<00:37,  2.23s/it]\n",
      "Loss: 0.01053:  93%|█████████▎| 233/250 [08:37<00:37,  2.23s/it]\n",
      "Loss: 0.00924:  93%|█████████▎| 233/250 [08:37<00:37,  2.23s/it]\n",
      "Loss: 0.01079:  93%|█████████▎| 233/250 [08:37<00:37,  2.23s/it]\n",
      "Loss: 0.00992:  93%|█████████▎| 233/250 [08:38<00:37,  2.23s/it]\n",
      "Loss: 0.01147:  93%|█████████▎| 233/250 [08:38<00:37,  2.23s/it]\n",
      "Loss: 0.01085:  93%|█████████▎| 233/250 [08:38<00:37,  2.23s/it]\n",
      "Loss: 0.01104:  93%|█████████▎| 233/250 [08:38<00:37,  2.23s/it]\n",
      "Loss: 0.00986:  93%|█████████▎| 233/250 [08:38<00:37,  2.23s/it]\n",
      "Loss: 0.00897:  93%|█████████▎| 233/250 [08:38<00:37,  2.23s/it]\n",
      "Loss: 0.00927:  93%|█████████▎| 233/250 [08:38<00:37,  2.23s/it]\n",
      "Loss: 0.00954:  93%|█████████▎| 233/250 [08:38<00:37,  2.23s/it]\n",
      "Loss: 0.01050:  93%|█████████▎| 233/250 [08:38<00:37,  2.23s/it]\n",
      "Loss: 0.00993:  93%|█████████▎| 233/250 [08:39<00:37,  2.23s/it]\n",
      "Loss: 0.01130:  93%|█████████▎| 233/250 [08:39<00:37,  2.23s/it]\n",
      "Loss: 0.01080:  93%|█████████▎| 233/250 [08:39<00:37,  2.23s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 54.47it/s]\u001b[A\n",
      "Loss: 0.01080:  94%|█████████▎| 234/250 [08:39<00:36,  2.25s/it]\n",
      "Loss: 0.01038:  94%|█████████▎| 234/250 [08:39<00:36,  2.25s/it]\n",
      "Loss: 0.01033:  94%|█████████▎| 234/250 [08:39<00:36,  2.25s/it]\n",
      "Loss: 0.01046:  94%|█████████▎| 234/250 [08:39<00:36,  2.25s/it]\n",
      "Loss: 0.01149:  94%|█████████▎| 234/250 [08:39<00:36,  2.25s/it]\n",
      "Loss: 0.01154:  94%|█████████▎| 234/250 [08:39<00:36,  2.25s/it]\n",
      "Loss: 0.01088:  94%|█████████▎| 234/250 [08:40<00:36,  2.25s/it]\n",
      "Loss: 0.01125:  94%|█████████▎| 234/250 [08:40<00:36,  2.25s/it]\n",
      "Loss: 0.01014:  94%|█████████▎| 234/250 [08:40<00:36,  2.25s/it]\n",
      "Loss: 0.00996:  94%|█████████▎| 234/250 [08:40<00:36,  2.25s/it]\n",
      "Loss: 0.00945:  94%|█████████▎| 234/250 [08:40<00:36,  2.25s/it]\n",
      "Loss: 0.01109:  94%|█████████▎| 234/250 [08:40<00:36,  2.25s/it]\n",
      "Loss: 0.01016:  94%|█████████▎| 234/250 [08:40<00:36,  2.25s/it]\n",
      "Loss: 0.00980:  94%|█████████▎| 234/250 [08:40<00:36,  2.25s/it]\n",
      "Loss: 0.00973:  94%|█████████▎| 234/250 [08:40<00:36,  2.25s/it]\n",
      "Loss: 0.01084:  94%|█████████▎| 234/250 [08:41<00:36,  2.25s/it]\n",
      "Loss: 0.00862:  94%|█████████▎| 234/250 [08:41<00:36,  2.25s/it]\n",
      "Loss: 0.01034:  94%|█████████▎| 234/250 [08:41<00:36,  2.25s/it]\n",
      "Loss: 0.01068:  94%|█████████▎| 234/250 [08:41<00:36,  2.25s/it]\n",
      "Loss: 0.00928:  94%|█████████▎| 234/250 [08:41<00:36,  2.25s/it]\n",
      "Loss: 0.01098:  94%|█████████▎| 234/250 [08:41<00:36,  2.25s/it]\n",
      "Loss: 0.01098:  94%|█████████▍| 235/250 [08:41<00:33,  2.24s/it]\n",
      "Loss: 0.01104:  94%|█████████▍| 235/250 [08:41<00:33,  2.24s/it]\n",
      "Loss: 0.01033:  94%|█████████▍| 235/250 [08:41<00:33,  2.24s/it]\n",
      "Loss: 0.01214:  94%|█████████▍| 235/250 [08:41<00:33,  2.24s/it]\n",
      "Loss: 0.00887:  94%|█████████▍| 235/250 [08:42<00:33,  2.24s/it]\n",
      "Loss: 0.01147:  94%|█████████▍| 235/250 [08:42<00:33,  2.24s/it]\n",
      "Loss: 0.01064:  94%|█████████▍| 235/250 [08:42<00:33,  2.24s/it]\n",
      "Loss: 0.01154:  94%|█████████▍| 235/250 [08:42<00:33,  2.24s/it]\n",
      "Loss: 0.01026:  94%|█████████▍| 235/250 [08:42<00:33,  2.24s/it]\n",
      "Loss: 0.00978:  94%|█████████▍| 235/250 [08:42<00:33,  2.24s/it]\n",
      "Loss: 0.01148:  94%|█████████▍| 235/250 [08:42<00:33,  2.24s/it]\n",
      "Loss: 0.01097:  94%|█████████▍| 235/250 [08:42<00:33,  2.24s/it]\n",
      "Loss: 0.01115:  94%|█████████▍| 235/250 [08:42<00:33,  2.24s/it]\n",
      "Loss: 0.00993:  94%|█████████▍| 235/250 [08:43<00:33,  2.24s/it]\n",
      "Loss: 0.00944:  94%|█████████▍| 235/250 [08:43<00:33,  2.24s/it]\n",
      "Loss: 0.00915:  94%|█████████▍| 235/250 [08:43<00:33,  2.24s/it]\n",
      "Loss: 0.00939:  94%|█████████▍| 235/250 [08:43<00:33,  2.24s/it]\n",
      "Loss: 0.01101:  94%|█████████▍| 235/250 [08:43<00:33,  2.24s/it]\n",
      "Loss: 0.01069:  94%|█████████▍| 235/250 [08:43<00:33,  2.24s/it]\n",
      "Loss: 0.00948:  94%|█████████▍| 235/250 [08:43<00:33,  2.24s/it]\n",
      "Loss: 0.01080:  94%|█████████▍| 235/250 [08:43<00:33,  2.24s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 53.99it/s]\u001b[A\n",
      "Loss: 0.01080:  94%|█████████▍| 236/250 [08:43<00:31,  2.24s/it]\n",
      "Loss: 0.00979:  94%|█████████▍| 236/250 [08:43<00:31,  2.24s/it]\n",
      "Loss: 0.01042:  94%|█████████▍| 236/250 [08:44<00:31,  2.24s/it]\n",
      "Loss: 0.01182:  94%|█████████▍| 236/250 [08:44<00:31,  2.24s/it]\n",
      "Loss: 0.00895:  94%|█████████▍| 236/250 [08:44<00:31,  2.24s/it]\n",
      "Loss: 0.01143:  94%|█████████▍| 236/250 [08:44<00:31,  2.24s/it]\n",
      "Loss: 0.01041:  94%|█████████▍| 236/250 [08:44<00:31,  2.24s/it]\n",
      "Loss: 0.00865:  94%|█████████▍| 236/250 [08:44<00:31,  2.24s/it]\n",
      "Loss: 0.01379:  94%|█████████▍| 236/250 [08:44<00:31,  2.24s/it]\n",
      "Loss: 0.00974:  94%|█████████▍| 236/250 [08:44<00:31,  2.24s/it]\n",
      "Loss: 0.01165:  94%|█████████▍| 236/250 [08:44<00:31,  2.24s/it]\n",
      "Loss: 0.00970:  94%|█████████▍| 236/250 [08:45<00:31,  2.24s/it]\n",
      "Loss: 0.00993:  94%|█████████▍| 236/250 [08:45<00:31,  2.24s/it]\n",
      "Loss: 0.01168:  94%|█████████▍| 236/250 [08:45<00:31,  2.24s/it]\n",
      "Loss: 0.01050:  94%|█████████▍| 236/250 [08:45<00:31,  2.24s/it]\n",
      "Loss: 0.01039:  94%|█████████▍| 236/250 [08:45<00:31,  2.24s/it]\n",
      "Loss: 0.01001:  94%|█████████▍| 236/250 [08:45<00:31,  2.24s/it]\n",
      "Loss: 0.01090:  94%|█████████▍| 236/250 [08:45<00:31,  2.24s/it]\n",
      "Loss: 0.01060:  94%|█████████▍| 236/250 [08:45<00:31,  2.24s/it]\n",
      "Loss: 0.01120:  94%|█████████▍| 236/250 [08:45<00:31,  2.24s/it]\n",
      "Loss: 0.01024:  94%|█████████▍| 236/250 [08:46<00:31,  2.24s/it]\n",
      "Loss: 0.01063:  94%|█████████▍| 236/250 [08:46<00:31,  2.24s/it]\n",
      "Loss: 0.01063:  95%|█████████▍| 237/250 [08:46<00:29,  2.24s/it]\n",
      "Loss: 0.01044:  95%|█████████▍| 237/250 [08:46<00:29,  2.24s/it]\n",
      "Loss: 0.01050:  95%|█████████▍| 237/250 [08:46<00:29,  2.24s/it]\n",
      "Loss: 0.01174:  95%|█████████▍| 237/250 [08:46<00:29,  2.24s/it]\n",
      "Loss: 0.00979:  95%|█████████▍| 237/250 [08:46<00:29,  2.24s/it]\n",
      "Loss: 0.01001:  95%|█████████▍| 237/250 [08:46<00:29,  2.24s/it]\n",
      "Loss: 0.01068:  95%|█████████▍| 237/250 [08:46<00:29,  2.24s/it]\n",
      "Loss: 0.01127:  95%|█████████▍| 237/250 [08:46<00:29,  2.24s/it]\n",
      "Loss: 0.01009:  95%|█████████▍| 237/250 [08:46<00:29,  2.24s/it]\n",
      "Loss: 0.01217:  95%|█████████▍| 237/250 [08:47<00:29,  2.24s/it]\n",
      "Loss: 0.01147:  95%|█████████▍| 237/250 [08:47<00:29,  2.24s/it]\n",
      "Loss: 0.01112:  95%|█████████▍| 237/250 [08:47<00:29,  2.24s/it]\n",
      "Loss: 0.01104:  95%|█████████▍| 237/250 [08:47<00:29,  2.24s/it]\n",
      "Loss: 0.00976:  95%|█████████▍| 237/250 [08:47<00:29,  2.24s/it]\n",
      "Loss: 0.00902:  95%|█████████▍| 237/250 [08:47<00:29,  2.24s/it]\n",
      "Loss: 0.00952:  95%|█████████▍| 237/250 [08:47<00:29,  2.24s/it]\n",
      "Loss: 0.00922:  95%|█████████▍| 237/250 [08:47<00:29,  2.24s/it]\n",
      "Loss: 0.01053:  95%|█████████▍| 237/250 [08:47<00:29,  2.24s/it]\n",
      "Loss: 0.01058:  95%|█████████▍| 237/250 [08:48<00:29,  2.24s/it]\n",
      "Loss: 0.00935:  95%|█████████▍| 237/250 [08:48<00:29,  2.24s/it]\n",
      "Loss: 0.01064:  95%|█████████▍| 237/250 [08:48<00:29,  2.24s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 53.96it/s]\u001b[A\n",
      "Loss: 0.01064:  95%|█████████▌| 238/250 [08:48<00:26,  2.24s/it]\n",
      "Loss: 0.01040:  95%|█████████▌| 238/250 [08:48<00:26,  2.24s/it]\n",
      "Loss: 0.01042:  95%|█████████▌| 238/250 [08:48<00:26,  2.24s/it]\n",
      "Loss: 0.01038:  95%|█████████▌| 238/250 [08:48<00:26,  2.24s/it]\n",
      "Loss: 0.01130:  95%|█████████▌| 238/250 [08:48<00:26,  2.24s/it]\n",
      "Loss: 0.01140:  95%|█████████▌| 238/250 [08:48<00:26,  2.24s/it]\n",
      "Loss: 0.01042:  95%|█████████▌| 238/250 [08:48<00:26,  2.24s/it]\n",
      "Loss: 0.01094:  95%|█████████▌| 238/250 [08:49<00:26,  2.24s/it]\n",
      "Loss: 0.01026:  95%|█████████▌| 238/250 [08:49<00:26,  2.24s/it]\n",
      "Loss: 0.01195:  95%|█████████▌| 238/250 [08:49<00:26,  2.24s/it]\n",
      "Loss: 0.01136:  95%|█████████▌| 238/250 [08:49<00:26,  2.24s/it]\n",
      "Loss: 0.01100:  95%|█████████▌| 238/250 [08:49<00:26,  2.24s/it]\n",
      "Loss: 0.01123:  95%|█████████▌| 238/250 [08:49<00:26,  2.24s/it]\n",
      "Loss: 0.01148:  95%|█████████▌| 238/250 [08:49<00:26,  2.24s/it]\n",
      "Loss: 0.01066:  95%|█████████▌| 238/250 [08:49<00:26,  2.24s/it]\n",
      "Loss: 0.01044:  95%|█████████▌| 238/250 [08:49<00:26,  2.24s/it]\n",
      "Loss: 0.01015:  95%|█████████▌| 238/250 [08:50<00:26,  2.24s/it]\n",
      "Loss: 0.01135:  95%|█████████▌| 238/250 [08:50<00:26,  2.24s/it]\n",
      "Loss: 0.00959:  95%|█████████▌| 238/250 [08:50<00:26,  2.24s/it]\n",
      "Loss: 0.01043:  95%|█████████▌| 238/250 [08:50<00:26,  2.24s/it]\n",
      "Loss: 0.01024:  95%|█████████▌| 238/250 [08:50<00:26,  2.24s/it]\n",
      "Loss: 0.01067:  95%|█████████▌| 238/250 [08:50<00:26,  2.24s/it]\n",
      "Loss: 0.01067:  96%|█████████▌| 239/250 [08:50<00:24,  2.25s/it]\n",
      "Loss: 0.00967:  96%|█████████▌| 239/250 [08:50<00:24,  2.25s/it]\n",
      "Loss: 0.01032:  96%|█████████▌| 239/250 [08:50<00:24,  2.25s/it]\n",
      "Loss: 0.01186:  96%|█████████▌| 239/250 [08:50<00:24,  2.25s/it]\n",
      "Loss: 0.00986:  96%|█████████▌| 239/250 [08:51<00:24,  2.25s/it]\n",
      "Loss: 0.01023:  96%|█████████▌| 239/250 [08:51<00:24,  2.25s/it]\n",
      "Loss: 0.01042:  96%|█████████▌| 239/250 [08:51<00:24,  2.25s/it]\n",
      "Loss: 0.01111:  96%|█████████▌| 239/250 [08:51<00:24,  2.25s/it]\n",
      "Loss: 0.01089:  96%|█████████▌| 239/250 [08:51<00:24,  2.25s/it]\n",
      "Loss: 0.00994:  96%|█████████▌| 239/250 [08:51<00:24,  2.25s/it]\n",
      "Loss: 0.00957:  96%|█████████▌| 239/250 [08:51<00:24,  2.25s/it]\n",
      "Loss: 0.01066:  96%|█████████▌| 239/250 [08:51<00:24,  2.25s/it]\n",
      "Loss: 0.00984:  96%|█████████▌| 239/250 [08:51<00:24,  2.25s/it]\n",
      "Loss: 0.00977:  96%|█████████▌| 239/250 [08:52<00:24,  2.25s/it]\n",
      "Loss: 0.00971:  96%|█████████▌| 239/250 [08:52<00:24,  2.25s/it]\n",
      "Loss: 0.00932:  96%|█████████▌| 239/250 [08:52<00:24,  2.25s/it]\n",
      "Loss: 0.00932:  96%|█████████▌| 239/250 [08:52<00:24,  2.25s/it]\n",
      "Loss: 0.01027:  96%|█████████▌| 239/250 [08:52<00:24,  2.25s/it]\n",
      "Loss: 0.00978:  96%|█████████▌| 239/250 [08:52<00:24,  2.25s/it]\n",
      "Loss: 0.01113:  96%|█████████▌| 239/250 [08:52<00:24,  2.25s/it]\n",
      "Loss: 0.01027:  96%|█████████▌| 239/250 [08:52<00:24,  2.25s/it]\n",
      "Loss: 0.01061:  96%|█████████▌| 239/250 [08:52<00:24,  2.25s/it]\n",
      "Loss: 0.01061:  96%|█████████▌| 240/250 [08:52<00:22,  2.25s/it]\n",
      "Loss: 0.00952:  96%|█████████▌| 240/250 [08:52<00:22,  2.25s/it]\n",
      "Loss: 0.01012:  96%|█████████▌| 240/250 [08:53<00:22,  2.25s/it]\n",
      "Loss: 0.01013:  96%|█████████▌| 240/250 [08:53<00:22,  2.25s/it]\n",
      "Loss: 0.00998:  96%|█████████▌| 240/250 [08:53<00:22,  2.25s/it]\n",
      "Loss: 0.01037:  96%|█████████▌| 240/250 [08:53<00:22,  2.25s/it]\n",
      "Loss: 0.01013:  96%|█████████▌| 240/250 [08:53<00:22,  2.25s/it]\n",
      "Loss: 0.00979:  96%|█████████▌| 240/250 [08:53<00:22,  2.25s/it]\n",
      "Loss: 0.01112:  96%|█████████▌| 240/250 [08:53<00:22,  2.25s/it]\n",
      "Loss: 0.01148:  96%|█████████▌| 240/250 [08:53<00:22,  2.25s/it]\n",
      "Loss: 0.01127:  96%|█████████▌| 240/250 [08:53<00:22,  2.25s/it]\n",
      "Loss: 0.00978:  96%|█████████▌| 240/250 [08:54<00:22,  2.25s/it]\n",
      "Loss: 0.01093:  96%|█████████▌| 240/250 [08:54<00:22,  2.25s/it]\n",
      "Loss: 0.01172:  96%|█████████▌| 240/250 [08:54<00:22,  2.25s/it]\n",
      "Loss: 0.00994:  96%|█████████▌| 240/250 [08:54<00:22,  2.25s/it]\n",
      "Loss: 0.01152:  96%|█████████▌| 240/250 [08:54<00:22,  2.25s/it]\n",
      "Loss: 0.00992:  96%|█████████▌| 240/250 [08:54<00:22,  2.25s/it]\n",
      "Loss: 0.01166:  96%|█████████▌| 240/250 [08:54<00:22,  2.25s/it]\n",
      "Loss: 0.01026:  96%|█████████▌| 240/250 [08:54<00:22,  2.25s/it]\n",
      "Loss: 0.01064:  96%|█████████▌| 240/250 [08:55<00:22,  2.25s/it]\n",
      "Loss: 0.01113:  96%|█████████▌| 240/250 [08:55<00:22,  2.25s/it]\n",
      "Loss: 0.01113:  96%|█████████▋| 241/250 [08:55<00:20,  2.24s/it]\n",
      "Loss: 0.00973:  96%|█████████▋| 241/250 [08:55<00:20,  2.24s/it]\n",
      "Loss: 0.01046:  96%|█████████▋| 241/250 [08:55<00:20,  2.24s/it]\n",
      "Loss: 0.01036:  96%|█████████▋| 241/250 [08:55<00:20,  2.24s/it]\n",
      "Loss: 0.00886:  96%|█████████▋| 241/250 [08:55<00:20,  2.24s/it]\n",
      "Loss: 0.00986:  96%|█████████▋| 241/250 [08:55<00:20,  2.24s/it]\n",
      "Loss: 0.00959:  96%|█████████▋| 241/250 [08:55<00:20,  2.24s/it]\n",
      "Loss: 0.00878:  96%|█████████▋| 241/250 [08:55<00:20,  2.24s/it]\n",
      "Loss: 0.01301:  96%|█████████▋| 241/250 [08:55<00:20,  2.24s/it]\n",
      "Loss: 0.00898:  96%|█████████▋| 241/250 [08:56<00:20,  2.24s/it]\n",
      "Loss: 0.01059:  96%|█████████▋| 241/250 [08:56<00:20,  2.24s/it]\n",
      "Loss: 0.01066:  96%|█████████▋| 241/250 [08:56<00:20,  2.24s/it]\n",
      "Loss: 0.00964:  96%|█████████▋| 241/250 [08:56<00:20,  2.24s/it]\n",
      "Loss: 0.01027:  96%|█████████▋| 241/250 [08:56<00:20,  2.24s/it]\n",
      "Loss: 0.01051:  96%|█████████▋| 241/250 [08:56<00:20,  2.24s/it]\n",
      "Loss: 0.01085:  96%|█████████▋| 241/250 [08:56<00:20,  2.24s/it]\n",
      "Loss: 0.00904:  96%|█████████▋| 241/250 [08:56<00:20,  2.24s/it]\n",
      "Loss: 0.01056:  96%|█████████▋| 241/250 [08:56<00:20,  2.24s/it]\n",
      "Loss: 0.01046:  96%|█████████▋| 241/250 [08:57<00:20,  2.24s/it]\n",
      "Loss: 0.00955:  96%|█████████▋| 241/250 [08:57<00:20,  2.24s/it]\n",
      "Loss: 0.00983:  96%|█████████▋| 241/250 [08:57<00:20,  2.24s/it]\n",
      "Loss: 0.01050:  96%|█████████▋| 241/250 [08:57<00:20,  2.24s/it]\n",
      "Loss: 0.01050:  97%|█████████▋| 242/250 [08:57<00:17,  2.25s/it]\n",
      "Loss: 0.00982:  97%|█████████▋| 242/250 [08:57<00:17,  2.25s/it]\n",
      "Loss: 0.01027:  97%|█████████▋| 242/250 [08:57<00:17,  2.25s/it]\n",
      "Loss: 0.01054:  97%|█████████▋| 242/250 [08:57<00:17,  2.25s/it]\n",
      "Loss: 0.01133:  97%|█████████▋| 242/250 [08:57<00:17,  2.25s/it]\n",
      "Loss: 0.01129:  97%|█████████▋| 242/250 [08:57<00:17,  2.25s/it]\n",
      "Loss: 0.01049:  97%|█████████▋| 242/250 [08:57<00:17,  2.25s/it]\n",
      "Loss: 0.01167:  97%|█████████▋| 242/250 [08:58<00:17,  2.25s/it]\n",
      "Loss: 0.00984:  97%|█████████▋| 242/250 [08:58<00:17,  2.25s/it]\n",
      "Loss: 0.01170:  97%|█████████▋| 242/250 [08:58<00:17,  2.25s/it]\n",
      "Loss: 0.01084:  97%|█████████▋| 242/250 [08:58<00:17,  2.25s/it]\n",
      "Loss: 0.00989:  97%|█████████▋| 242/250 [08:58<00:17,  2.25s/it]\n",
      "Loss: 0.00966:  97%|█████████▋| 242/250 [08:58<00:17,  2.25s/it]\n",
      "Loss: 0.01041:  97%|█████████▋| 242/250 [08:58<00:17,  2.25s/it]\n",
      "Loss: 0.00892:  97%|█████████▋| 242/250 [08:58<00:17,  2.25s/it]\n",
      "Loss: 0.00950:  97%|█████████▋| 242/250 [08:58<00:17,  2.25s/it]\n",
      "Loss: 0.01124:  97%|█████████▋| 242/250 [08:59<00:17,  2.25s/it]\n",
      "Loss: 0.01112:  97%|█████████▋| 242/250 [08:59<00:17,  2.25s/it]\n",
      "Loss: 0.00978:  97%|█████████▋| 242/250 [08:59<00:17,  2.25s/it]\n",
      "Loss: 0.01031:  97%|█████████▋| 242/250 [08:59<00:17,  2.25s/it]\n",
      "Loss: 0.01048:  97%|█████████▋| 242/250 [08:59<00:17,  2.25s/it]\n",
      "Loss: 0.01108:  97%|█████████▋| 242/250 [08:59<00:17,  2.25s/it]\n",
      "Loss: 0.01108:  97%|█████████▋| 243/250 [08:59<00:15,  2.24s/it]\n",
      "Loss: 0.00996:  97%|█████████▋| 243/250 [08:59<00:15,  2.24s/it]\n",
      "Loss: 0.01021:  97%|█████████▋| 243/250 [08:59<00:15,  2.24s/it]\n",
      "Loss: 0.01207:  97%|█████████▋| 243/250 [08:59<00:15,  2.24s/it]\n",
      "Loss: 0.00885:  97%|█████████▋| 243/250 [08:59<00:15,  2.24s/it]\n",
      "Loss: 0.00928:  97%|█████████▋| 243/250 [09:00<00:15,  2.24s/it]\n",
      "Loss: 0.01066:  97%|█████████▋| 243/250 [09:00<00:15,  2.24s/it]\n",
      "Loss: 0.00847:  97%|█████████▋| 243/250 [09:00<00:15,  2.24s/it]\n",
      "Loss: 0.01318:  97%|█████████▋| 243/250 [09:00<00:15,  2.24s/it]\n",
      "Loss: 0.00921:  97%|█████████▋| 243/250 [09:00<00:15,  2.24s/it]\n",
      "Loss: 0.01173:  97%|█████████▋| 243/250 [09:00<00:15,  2.24s/it]\n",
      "Loss: 0.01172:  97%|█████████▋| 243/250 [09:00<00:15,  2.24s/it]\n",
      "Loss: 0.00990:  97%|█████████▋| 243/250 [09:00<00:15,  2.24s/it]\n",
      "Loss: 0.01115:  97%|█████████▋| 243/250 [09:00<00:15,  2.24s/it]\n",
      "Loss: 0.01241:  97%|█████████▋| 243/250 [09:01<00:15,  2.24s/it]\n",
      "Loss: 0.01015:  97%|█████████▋| 243/250 [09:01<00:15,  2.24s/it]\n",
      "Loss: 0.01105:  97%|█████████▋| 243/250 [09:01<00:15,  2.24s/it]\n",
      "Loss: 0.00937:  97%|█████████▋| 243/250 [09:01<00:15,  2.24s/it]\n",
      "Loss: 0.01147:  97%|█████████▋| 243/250 [09:01<00:15,  2.24s/it]\n",
      "Loss: 0.01127:  97%|█████████▋| 243/250 [09:01<00:15,  2.24s/it]\n",
      "Loss: 0.01026:  97%|█████████▋| 243/250 [09:01<00:15,  2.24s/it]\n",
      "Loss: 0.01037:  97%|█████████▋| 243/250 [09:01<00:15,  2.24s/it]\n",
      "Loss: 0.01037:  98%|█████████▊| 244/250 [09:01<00:13,  2.24s/it]\n",
      "Loss: 0.00965:  98%|█████████▊| 244/250 [09:01<00:13,  2.24s/it]\n",
      "Loss: 0.01032:  98%|█████████▊| 244/250 [09:02<00:13,  2.24s/it]\n",
      "Loss: 0.01187:  98%|█████████▊| 244/250 [09:02<00:13,  2.24s/it]\n",
      "Loss: 0.00880:  98%|█████████▊| 244/250 [09:02<00:13,  2.24s/it]\n",
      "Loss: 0.00896:  98%|█████████▊| 244/250 [09:02<00:13,  2.24s/it]\n",
      "Loss: 0.01048:  98%|█████████▊| 244/250 [09:02<00:13,  2.24s/it]\n",
      "Loss: 0.00838:  98%|█████████▊| 244/250 [09:02<00:13,  2.24s/it]\n",
      "Loss: 0.01361:  98%|█████████▊| 244/250 [09:02<00:13,  2.24s/it]\n",
      "Loss: 0.00900:  98%|█████████▊| 244/250 [09:02<00:13,  2.24s/it]\n",
      "Loss: 0.01125:  98%|█████████▊| 244/250 [09:02<00:13,  2.24s/it]\n",
      "Loss: 0.01175:  98%|█████████▊| 244/250 [09:02<00:13,  2.24s/it]\n",
      "Loss: 0.00964:  98%|█████████▊| 244/250 [09:03<00:13,  2.24s/it]\n",
      "Loss: 0.01022:  98%|█████████▊| 244/250 [09:03<00:13,  2.24s/it]\n",
      "Loss: 0.00913:  98%|█████████▊| 244/250 [09:03<00:13,  2.24s/it]\n",
      "Loss: 0.00994:  98%|█████████▊| 244/250 [09:03<00:13,  2.24s/it]\n",
      "Loss: 0.01132:  98%|█████████▊| 244/250 [09:03<00:13,  2.24s/it]\n",
      "Loss: 0.00946:  98%|█████████▊| 244/250 [09:03<00:13,  2.24s/it]\n",
      "Loss: 0.01171:  98%|█████████▊| 244/250 [09:03<00:13,  2.24s/it]\n",
      "Loss: 0.01031:  98%|█████████▊| 244/250 [09:03<00:13,  2.24s/it]\n",
      "Loss: 0.01056:  98%|█████████▊| 244/250 [09:03<00:13,  2.24s/it]\n",
      "Loss: 0.01060:  98%|█████████▊| 244/250 [09:03<00:13,  2.24s/it]\n",
      "Loss: 0.01060:  98%|█████████▊| 245/250 [09:03<00:11,  2.23s/it]\n",
      "Loss: 0.01024:  98%|█████████▊| 245/250 [09:04<00:11,  2.23s/it]\n",
      "Loss: 0.01041:  98%|█████████▊| 245/250 [09:04<00:11,  2.23s/it]\n",
      "Loss: 0.01179:  98%|█████████▊| 245/250 [09:04<00:11,  2.23s/it]\n",
      "Loss: 0.00895:  98%|█████████▊| 245/250 [09:04<00:11,  2.23s/it]\n",
      "Loss: 0.00897:  98%|█████████▊| 245/250 [09:04<00:11,  2.23s/it]\n",
      "Loss: 0.01087:  98%|█████████▊| 245/250 [09:04<00:11,  2.23s/it]\n",
      "Loss: 0.00824:  98%|█████████▊| 245/250 [09:04<00:11,  2.23s/it]\n",
      "Loss: 0.01296:  98%|█████████▊| 245/250 [09:04<00:11,  2.23s/it]\n",
      "Loss: 0.00919:  98%|█████████▊| 245/250 [09:04<00:11,  2.23s/it]\n",
      "Loss: 0.01113:  98%|█████████▊| 245/250 [09:05<00:11,  2.23s/it]\n",
      "Loss: 0.01182:  98%|█████████▊| 245/250 [09:05<00:11,  2.23s/it]\n",
      "Loss: 0.01022:  98%|█████████▊| 245/250 [09:05<00:11,  2.23s/it]\n",
      "Loss: 0.01069:  98%|█████████▊| 245/250 [09:05<00:11,  2.23s/it]\n",
      "Loss: 0.01190:  98%|█████████▊| 245/250 [09:05<00:11,  2.23s/it]\n",
      "Loss: 0.00938:  98%|█████████▊| 245/250 [09:05<00:11,  2.23s/it]\n",
      "Loss: 0.00925:  98%|█████████▊| 245/250 [09:05<00:11,  2.23s/it]\n",
      "Loss: 0.01024:  98%|█████████▊| 245/250 [09:05<00:11,  2.23s/it]\n",
      "Loss: 0.01032:  98%|█████████▊| 245/250 [09:05<00:11,  2.23s/it]\n",
      "Loss: 0.01229:  98%|█████████▊| 245/250 [09:06<00:11,  2.23s/it]\n",
      "Loss: 0.01043:  98%|█████████▊| 245/250 [09:06<00:11,  2.23s/it]\n",
      "Loss: 0.01053:  98%|█████████▊| 245/250 [09:06<00:11,  2.23s/it]\n",
      "Loss: 0.01053:  98%|█████████▊| 246/250 [09:06<00:08,  2.23s/it]\n",
      "Loss: 0.01052:  98%|█████████▊| 246/250 [09:06<00:08,  2.23s/it]\n",
      "Loss: 0.01012:  98%|█████████▊| 246/250 [09:06<00:08,  2.23s/it]\n",
      "Loss: 0.01003:  98%|█████████▊| 246/250 [09:06<00:08,  2.23s/it]\n",
      "Loss: 0.01143:  98%|█████████▊| 246/250 [09:06<00:08,  2.23s/it]\n",
      "Loss: 0.01153:  98%|█████████▊| 246/250 [09:06<00:08,  2.23s/it]\n",
      "Loss: 0.01037:  98%|█████████▊| 246/250 [09:06<00:08,  2.23s/it]\n",
      "Loss: 0.01116:  98%|█████████▊| 246/250 [09:06<00:08,  2.23s/it]\n",
      "Loss: 0.01011:  98%|█████████▊| 246/250 [09:07<00:08,  2.23s/it]\n",
      "Loss: 0.01178:  98%|█████████▊| 246/250 [09:07<00:08,  2.23s/it]\n",
      "Loss: 0.01070:  98%|█████████▊| 246/250 [09:07<00:08,  2.23s/it]\n",
      "Loss: 0.01098:  98%|█████████▊| 246/250 [09:07<00:08,  2.23s/it]\n",
      "Loss: 0.01096:  98%|█████████▊| 246/250 [09:07<00:08,  2.23s/it]\n",
      "Loss: 0.01139:  98%|█████████▊| 246/250 [09:07<00:08,  2.23s/it]\n",
      "Loss: 0.01047:  98%|█████████▊| 246/250 [09:07<00:08,  2.23s/it]\n",
      "Loss: 0.01041:  98%|█████████▊| 246/250 [09:07<00:08,  2.23s/it]\n",
      "Loss: 0.00999:  98%|█████████▊| 246/250 [09:07<00:08,  2.23s/it]\n",
      "Loss: 0.01119:  98%|█████████▊| 246/250 [09:08<00:08,  2.23s/it]\n",
      "Loss: 0.00977:  98%|█████████▊| 246/250 [09:08<00:08,  2.23s/it]\n",
      "Loss: 0.01097:  98%|█████████▊| 246/250 [09:08<00:08,  2.23s/it]\n",
      "Loss: 0.01043:  98%|█████████▊| 246/250 [09:08<00:08,  2.23s/it]\n",
      "Loss: 0.01062:  98%|█████████▊| 246/250 [09:08<00:08,  2.23s/it]\n",
      "Loss: 0.01062:  99%|█████████▉| 247/250 [09:08<00:06,  2.22s/it]\n",
      "Loss: 0.01007:  99%|█████████▉| 247/250 [09:08<00:06,  2.22s/it]\n",
      "Loss: 0.01031:  99%|█████████▉| 247/250 [09:08<00:06,  2.22s/it]\n",
      "Loss: 0.01222:  99%|█████████▉| 247/250 [09:08<00:06,  2.22s/it]\n",
      "Loss: 0.00882:  99%|█████████▉| 247/250 [09:08<00:06,  2.22s/it]\n",
      "Loss: 0.00917:  99%|█████████▉| 247/250 [09:08<00:06,  2.22s/it]\n",
      "Loss: 0.01081:  99%|█████████▉| 247/250 [09:09<00:06,  2.22s/it]\n",
      "Loss: 0.00822:  99%|█████████▉| 247/250 [09:09<00:06,  2.22s/it]\n",
      "Loss: 0.01296:  99%|█████████▉| 247/250 [09:09<00:06,  2.22s/it]\n",
      "Loss: 0.00907:  99%|█████████▉| 247/250 [09:09<00:06,  2.22s/it]\n",
      "Loss: 0.01132:  99%|█████████▉| 247/250 [09:09<00:06,  2.22s/it]\n",
      "Loss: 0.01138:  99%|█████████▉| 247/250 [09:09<00:06,  2.22s/it]\n",
      "Loss: 0.01004:  99%|█████████▉| 247/250 [09:09<00:06,  2.22s/it]\n",
      "Loss: 0.01077:  99%|█████████▉| 247/250 [09:09<00:06,  2.22s/it]\n",
      "Loss: 0.01160:  99%|█████████▉| 247/250 [09:09<00:06,  2.22s/it]\n",
      "Loss: 0.01025:  99%|█████████▉| 247/250 [09:10<00:06,  2.22s/it]\n",
      "Loss: 0.00955:  99%|█████████▉| 247/250 [09:10<00:06,  2.22s/it]\n",
      "Loss: 0.01036:  99%|█████████▉| 247/250 [09:10<00:06,  2.22s/it]\n",
      "Loss: 0.01082:  99%|█████████▉| 247/250 [09:10<00:06,  2.22s/it]\n",
      "Loss: 0.01244:  99%|█████████▉| 247/250 [09:10<00:06,  2.22s/it]\n",
      "Loss: 0.01028:  99%|█████████▉| 247/250 [09:10<00:06,  2.22s/it]\n",
      "Loss: 0.01067:  99%|█████████▉| 247/250 [09:10<00:06,  2.22s/it]\n",
      "Loss: 0.01067:  99%|█████████▉| 248/250 [09:10<00:04,  2.22s/it]\n",
      "Loss: 0.00973:  99%|█████████▉| 248/250 [09:10<00:04,  2.22s/it]\n",
      "Loss: 0.01043:  99%|█████████▉| 248/250 [09:10<00:04,  2.22s/it]\n",
      "Loss: 0.01185:  99%|█████████▉| 248/250 [09:10<00:04,  2.22s/it]\n",
      "Loss: 0.00886:  99%|█████████▉| 248/250 [09:11<00:04,  2.22s/it]\n",
      "Loss: 0.00877:  99%|█████████▉| 248/250 [09:11<00:04,  2.22s/it]\n",
      "Loss: 0.01090:  99%|█████████▉| 248/250 [09:11<00:04,  2.22s/it]\n",
      "Loss: 0.00816:  99%|█████████▉| 248/250 [09:11<00:04,  2.22s/it]\n",
      "Loss: 0.01277:  99%|█████████▉| 248/250 [09:11<00:04,  2.22s/it]\n",
      "Loss: 0.00918:  99%|█████████▉| 248/250 [09:11<00:04,  2.22s/it]\n",
      "Loss: 0.01155:  99%|█████████▉| 248/250 [09:11<00:04,  2.22s/it]\n",
      "Loss: 0.01184:  99%|█████████▉| 248/250 [09:11<00:04,  2.22s/it]\n",
      "Loss: 0.00988:  99%|█████████▉| 248/250 [09:11<00:04,  2.22s/it]\n",
      "Loss: 0.01099:  99%|█████████▉| 248/250 [09:12<00:04,  2.22s/it]\n",
      "Loss: 0.01178:  99%|█████████▉| 248/250 [09:12<00:04,  2.22s/it]\n",
      "Loss: 0.00953:  99%|█████████▉| 248/250 [09:12<00:04,  2.22s/it]\n",
      "Loss: 0.00944:  99%|█████████▉| 248/250 [09:12<00:04,  2.22s/it]\n",
      "Loss: 0.01062:  99%|█████████▉| 248/250 [09:12<00:04,  2.22s/it]\n",
      "Loss: 0.01065:  99%|█████████▉| 248/250 [09:12<00:04,  2.22s/it]\n",
      "Loss: 0.01215:  99%|█████████▉| 248/250 [09:12<00:04,  2.22s/it]\n",
      "Loss: 0.01022:  99%|█████████▉| 248/250 [09:12<00:04,  2.22s/it]\n",
      "Loss: 0.01040:  99%|█████████▉| 248/250 [09:12<00:04,  2.22s/it]\n",
      "Loss: 0.01040: 100%|█████████▉| 249/250 [09:12<00:02,  2.22s/it]\n",
      "Loss: 0.00932: 100%|█████████▉| 249/250 [09:12<00:02,  2.22s/it]\n",
      "Loss: 0.01041: 100%|█████████▉| 249/250 [09:13<00:02,  2.22s/it]\n",
      "Loss: 0.01176: 100%|█████████▉| 249/250 [09:13<00:02,  2.22s/it]\n",
      "Loss: 0.00888: 100%|█████████▉| 249/250 [09:13<00:02,  2.22s/it]\n",
      "Loss: 0.00893: 100%|█████████▉| 249/250 [09:13<00:02,  2.22s/it]\n",
      "Loss: 0.01062: 100%|█████████▉| 249/250 [09:13<00:02,  2.22s/it]\n",
      "Loss: 0.00825: 100%|█████████▉| 249/250 [09:13<00:02,  2.22s/it]\n",
      "Loss: 0.01320: 100%|█████████▉| 249/250 [09:13<00:02,  2.22s/it]\n",
      "Loss: 0.00896: 100%|█████████▉| 249/250 [09:13<00:02,  2.22s/it]\n",
      "Loss: 0.01132: 100%|█████████▉| 249/250 [09:13<00:02,  2.22s/it]\n",
      "Loss: 0.01181: 100%|█████████▉| 249/250 [09:14<00:02,  2.22s/it]\n",
      "Loss: 0.00990: 100%|█████████▉| 249/250 [09:14<00:02,  2.22s/it]\n",
      "Loss: 0.01084: 100%|█████████▉| 249/250 [09:14<00:02,  2.22s/it]\n",
      "Loss: 0.01150: 100%|█████████▉| 249/250 [09:14<00:02,  2.22s/it]\n",
      "Loss: 0.00988: 100%|█████████▉| 249/250 [09:14<00:02,  2.22s/it]\n",
      "Loss: 0.00934: 100%|█████████▉| 249/250 [09:14<00:02,  2.22s/it]\n",
      "Loss: 0.01008: 100%|█████████▉| 249/250 [09:14<00:02,  2.22s/it]\n",
      "Loss: 0.01040: 100%|█████████▉| 249/250 [09:14<00:02,  2.22s/it]\n",
      "Loss: 0.01327: 100%|█████████▉| 249/250 [09:14<00:02,  2.22s/it]\n",
      "Loss: 0.01026: 100%|█████████▉| 249/250 [09:15<00:02,  2.22s/it]\n",
      "Loss: 0.01059: 100%|█████████▉| 249/250 [09:15<00:02,  2.22s/it]\n"
     ]
    }
   ],
   "source": [
    "from PIL import Image\n",
    "n_epochs = 250\n",
    "with tf.Session() as sess:\n",
    "    train_writer = tf.summary.FileWriter('./log/{}'.format(current_date), sess.graph)\n",
    "    sess.run(init)  # run the initializer\n",
    "\n",
    "    res = sess.run([*trn_summary], feed_dict={X: data_test[:25]})\n",
    "\n",
    "    train_writer.add_summary(res[1], -1)\n",
    "    for i in range(n_distribution):\n",
    "        train_writer.add_summary(res[2][i], -1)\n",
    "        train_writer.add_summary(res[3][i], -1)\n",
    "    train_writer.add_summary(res[4], -1)\n",
    "\n",
    "    epoch_tqdm = tqdm(range(1, n_epochs + 1), desc=\"Loss\", leave=False)\n",
    "    for epoch in epoch_tqdm:\n",
    "        n_batch = data_train.shape[0] // batch_size\n",
    "        for iteration in tqdm(range(n_batch), desc=\"Batches\", leave=False):\n",
    "            batch_x = data_train[(iteration * batch_size):((iteration + 1) * batch_size), :]\n",
    "\n",
    "            # Run optimization op (backprop) and cost op (to get loss value)\n",
    "            res = sess.run([optimizer, loss, *trn_summary, *trn_imgs], feed_dict={X: batch_x})\n",
    "\n",
    "            train_writer.add_summary(res[-2], n_batch * (epoch - 1) + iteration)\n",
    "            train_writer.add_summary(res[-1], n_batch * (epoch - 1) + iteration)\n",
    "            train_writer.add_summary(res[2], n_batch * (epoch - 1) + iteration)\n",
    "            train_writer.add_summary(res[3], n_batch * (epoch - 1) + iteration)\n",
    "            for i in range(n_distribution):\n",
    "                train_writer.add_summary(res[4][i], n_batch * (epoch - 1) + iteration)\n",
    "                train_writer.add_summary(res[5][i], n_batch * (epoch - 1) + iteration)\n",
    "            train_writer.add_summary(res[6], n_batch * (epoch - 1) + iteration)\n",
    "\n",
    "            epoch_tqdm.set_description(\"Loss: {:.5f}\".format(res[1]))\n",
    "\n",
    "        tst_loss, tst_input, tst_output = sess.run([*tst_summary], feed_dict={X: data_test[:25]})\n",
    "        train_writer.add_summary(tst_loss, epoch)\n",
    "        train_writer.add_summary(tst_input, epoch)\n",
    "        train_writer.add_summary(tst_output, epoch)\n",
    "    out_train = sess.run([decoder_op, loss], feed_dict={X: data_train})\n",
    "    out_test = sess.run([decoder_op, loss], feed_dict={X: data_test})\n",
    "    for i, image in enumerate(out_train[0].reshape(8000, 28, 28)*255):\n",
    "        pil_image = Image.fromarray(image)\n",
    "        pil_image.convert('L').save('{}/imgAEtrain/{}.jpg'.format(ds_label,i))\n",
    "    for i, image in enumerate(out_test[0].reshape(2000, 28, 28)*255):\n",
    "        pil_image = Image.fromarray(image)\n",
    "        pil_image.convert('L').save('{}/imgAEtest/{}.jpg'.format(ds_label,i+8000))\n",
    "    result = open('{}/res.txt'.format(ds_label), 'w')\n",
    "    result.write(str(out_train[1])+'\\n')\n",
    "    result.write(str(out_test[1]))\n",
    "    result.close()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  },
  "pycharm": {
   "stem_cell": {
    "cell_type": "raw",
    "metadata": {
     "collapsed": false
    },
    "source": []
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}