aeCIFAR.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": 7,
   "metadata": {},
   "outputs": [],
   "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",
    "\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 = 32*32\n",
    "\n",
    "n_distribution = 5  # number of n_distribution\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([32 * 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], 14)] = 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)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Read data done in 1.115s.\n",
      "GMM done in 6.772s.\n"
     ]
    }
   ],
   "source": [
    "t0 = time()\n",
    "images = list()\n",
    "for i in range(10_000):\n",
    "    img: Image = Image.open('CIFAR/img/{}.jpg'.format(i))\n",
    "    images.append(np.array(img))\n",
    "images = np.array(images)\n",
    "labels = list()\n",
    "masks = list()\n",
    "labels_file = open('labelsCIFAR.csv', '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[:8000])\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",
    "\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), (32, 32), 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), (32, 32), 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": 9,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Loss:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.09879:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.09228:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.09687:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.10147:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.09773:   0%|          | 0/250 [00:00<?, ?it/s]\n",
      "Loss: 0.09733:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.09074:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.09243:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.09430:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.09610:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.09463:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.08858:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.09223:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.10306:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.09434:   0%|          | 0/250 [00:01<?, ?it/s]\n",
      "Loss: 0.08771:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.08283:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.07272:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.07598:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.06537:   0%|          | 0/250 [00:02<?, ?it/s]\n",
      "Loss: 0.06537:   0%|          | 1/250 [00:02<10:58,  2.64s/it]\n",
      "Loss: 0.06261:   0%|          | 1/250 [00:02<10:58,  2.64s/it]\n",
      "Loss: 0.05695:   0%|          | 1/250 [00:02<10:58,  2.64s/it]\n",
      "Loss: 0.05285:   0%|          | 1/250 [00:02<10:58,  2.64s/it]\n",
      "Loss: 0.04805:   0%|          | 1/250 [00:03<10:58,  2.64s/it]\n",
      "Loss: 0.04601:   0%|          | 1/250 [00:03<10:58,  2.64s/it]\n",
      "Loss: 0.04429:   0%|          | 1/250 [00:03<10:58,  2.64s/it]\n",
      "Loss: 0.04757:   0%|          | 1/250 [00:03<10:58,  2.64s/it]\n",
      "Loss: 0.05012:   0%|          | 1/250 [00:03<10:58,  2.64s/it]\n",
      "Loss: 0.04592:   0%|          | 1/250 [00:03<10:58,  2.64s/it]\n",
      "Loss: 0.05358:   0%|          | 1/250 [00:03<10:58,  2.64s/it]\n",
      "Loss: 0.04913:   0%|          | 1/250 [00:03<10:58,  2.64s/it]\n",
      "Loss: 0.05070:   0%|          | 1/250 [00:03<10:58,  2.64s/it]\n",
      "Loss: 0.05142:   0%|          | 1/250 [00:04<10:58,  2.64s/it]\n",
      "Loss: 0.05188:   0%|          | 1/250 [00:04<10:58,  2.64s/it]\n",
      "Loss: 0.04689:   0%|          | 1/250 [00:04<10:58,  2.64s/it]\n",
      "Loss: 0.04232:   0%|          | 1/250 [00:04<10:58,  2.64s/it]\n",
      "Loss: 0.05310:   0%|          | 1/250 [00:04<10:58,  2.64s/it]\n",
      "Loss: 0.04287:   0%|          | 1/250 [00:04<10:58,  2.64s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 62.45it/s]\u001b[A\n",
      "Loss: 0.04287:   1%|          | 2/250 [00:04<10:07,  2.45s/it]\n",
      "Loss: 0.04756:   1%|          | 2/250 [00:04<10:07,  2.45s/it]\n",
      "Loss: 0.05731:   1%|          | 2/250 [00:04<10:07,  2.45s/it]\n",
      "Loss: 0.04587:   1%|          | 2/250 [00:04<10:07,  2.45s/it]\n",
      "Loss: 0.04338:   1%|          | 2/250 [00:05<10:07,  2.45s/it]\n",
      "Loss: 0.03868:   1%|          | 2/250 [00:05<10:07,  2.45s/it]\n",
      "Loss: 0.03959:   1%|          | 2/250 [00:05<10:07,  2.45s/it]\n",
      "Loss: 0.03647:   1%|          | 2/250 [00:05<10:07,  2.45s/it]\n",
      "Loss: 0.04900:   1%|          | 2/250 [00:05<10:07,  2.45s/it]\n",
      "Loss: 0.04043:   1%|          | 2/250 [00:05<10:07,  2.45s/it]\n",
      "Loss: 0.03543:   1%|          | 2/250 [00:05<10:07,  2.45s/it]\n",
      "Loss: 0.03448:   1%|          | 2/250 [00:05<10:07,  2.45s/it]\n",
      "Loss: 0.03987:   1%|          | 2/250 [00:05<10:07,  2.45s/it]\n",
      "Loss: 0.04103:   1%|          | 2/250 [00:06<10:07,  2.45s/it]\n",
      "Loss: 0.04074:   1%|          | 2/250 [00:06<10:07,  2.45s/it]\n",
      "Loss: 0.03630:   1%|          | 2/250 [00:06<10:07,  2.45s/it]\n",
      "Loss: 0.03549:   1%|          | 2/250 [00:06<10:07,  2.45s/it]\n",
      "Loss: 0.03497:   1%|          | 2/250 [00:06<10:07,  2.45s/it]\n",
      "Loss: 0.04330:   1%|          | 2/250 [00:06<10:07,  2.45s/it]\n",
      "Loss: 0.03185:   1%|          | 2/250 [00:06<10:07,  2.45s/it]\n",
      "Loss: 0.03185:   1%|          | 3/250 [00:06<09:34,  2.32s/it]\n",
      "Loss: 0.04206:   1%|          | 3/250 [00:06<09:34,  2.32s/it]\n",
      "Loss: 0.03363:   1%|          | 3/250 [00:06<09:34,  2.32s/it]\n",
      "Loss: 0.04144:   1%|          | 3/250 [00:07<09:34,  2.32s/it]\n",
      "Loss: 0.03464:   1%|          | 3/250 [00:07<09:34,  2.32s/it]\n",
      "Loss: 0.04514:   1%|          | 3/250 [00:07<09:34,  2.32s/it]\n",
      "Loss: 0.03147:   1%|          | 3/250 [00:07<09:34,  2.32s/it]\n",
      "Loss: 0.03735:   1%|          | 3/250 [00:07<09:34,  2.32s/it]\n",
      "Loss: 0.03548:   1%|          | 3/250 [00:07<09:34,  2.32s/it]\n",
      "Loss: 0.03655:   1%|          | 3/250 [00:07<09:34,  2.32s/it]\n",
      "Loss: 0.03844:   1%|          | 3/250 [00:07<09:34,  2.32s/it]\n",
      "Loss: 0.03353:   1%|          | 3/250 [00:07<09:34,  2.32s/it]\n",
      "Loss: 0.03859:   1%|          | 3/250 [00:08<09:34,  2.32s/it]\n",
      "Loss: 0.03517:   1%|          | 3/250 [00:08<09:34,  2.32s/it]\n",
      "Loss: 0.03518:   1%|          | 3/250 [00:08<09:34,  2.32s/it]\n",
      "Loss: 0.03352:   1%|          | 3/250 [00:08<09:34,  2.32s/it]\n",
      "Loss: 0.03766:   1%|          | 3/250 [00:08<09:34,  2.32s/it]\n",
      "Loss: 0.03518:   1%|          | 3/250 [00:08<09:34,  2.32s/it]\n",
      "Loss: 0.03158:   1%|          | 3/250 [00:08<09:34,  2.32s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.57it/s]\u001b[A\n",
      "Loss: 0.03158:   2%|▏         | 4/250 [00:08<09:11,  2.24s/it]\n",
      "Loss: 0.03359:   2%|▏         | 4/250 [00:08<09:11,  2.24s/it]\n",
      "Loss: 0.04137:   2%|▏         | 4/250 [00:08<09:11,  2.24s/it]\n",
      "Loss: 0.03554:   2%|▏         | 4/250 [00:09<09:11,  2.24s/it]\n",
      "Loss: 0.03573:   2%|▏         | 4/250 [00:09<09:11,  2.24s/it]\n",
      "Loss: 0.03249:   2%|▏         | 4/250 [00:09<09:11,  2.24s/it]\n",
      "Loss: 0.03530:   2%|▏         | 4/250 [00:09<09:11,  2.24s/it]\n",
      "Loss: 0.03321:   2%|▏         | 4/250 [00:09<09:11,  2.24s/it]\n",
      "Loss: 0.03995:   2%|▏         | 4/250 [00:09<09:11,  2.24s/it]\n",
      "Loss: 0.03357:   2%|▏         | 4/250 [00:09<09:11,  2.24s/it]\n",
      "Loss: 0.03789:   2%|▏         | 4/250 [00:09<09:11,  2.24s/it]\n",
      "Loss: 0.03249:   2%|▏         | 4/250 [00:09<09:11,  2.24s/it]\n",
      "Loss: 0.03732:   2%|▏         | 4/250 [00:10<09:11,  2.24s/it]\n",
      "Loss: 0.03110:   2%|▏         | 4/250 [00:10<09:11,  2.24s/it]\n",
      "Loss: 0.03552:   2%|▏         | 4/250 [00:10<09:11,  2.24s/it]\n",
      "Loss: 0.03103:   2%|▏         | 4/250 [00:10<09:11,  2.24s/it]\n",
      "Loss: 0.03930:   2%|▏         | 4/250 [00:10<09:11,  2.24s/it]\n",
      "Loss: 0.03441:   2%|▏         | 4/250 [00:10<09:11,  2.24s/it]\n",
      "Loss: 0.03006:   2%|▏         | 4/250 [00:10<09:11,  2.24s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 60.18it/s]\u001b[A\n",
      "Loss: 0.03006:   2%|▏         | 5/250 [00:10<08:54,  2.18s/it]\n",
      "Loss: 0.03284:   2%|▏         | 5/250 [00:10<08:54,  2.18s/it]\n",
      "Loss: 0.03478:   2%|▏         | 5/250 [00:10<08:54,  2.18s/it]\n",
      "Loss: 0.03502:   2%|▏         | 5/250 [00:11<08:54,  2.18s/it]\n",
      "Loss: 0.03161:   2%|▏         | 5/250 [00:11<08:54,  2.18s/it]\n",
      "Loss: 0.03344:   2%|▏         | 5/250 [00:11<08:54,  2.18s/it]\n",
      "Loss: 0.03113:   2%|▏         | 5/250 [00:11<08:54,  2.18s/it]\n",
      "Loss: 0.03314:   2%|▏         | 5/250 [00:11<08:54,  2.18s/it]\n",
      "Loss: 0.03206:   2%|▏         | 5/250 [00:11<08:54,  2.18s/it]\n",
      "Loss: 0.03484:   2%|▏         | 5/250 [00:11<08:54,  2.18s/it]\n",
      "Loss: 0.03304:   2%|▏         | 5/250 [00:11<08:54,  2.18s/it]\n",
      "Loss: 0.03318:   2%|▏         | 5/250 [00:11<08:54,  2.18s/it]\n",
      "Loss: 0.03086:   2%|▏         | 5/250 [00:12<08:54,  2.18s/it]\n",
      "Loss: 0.03517:   2%|▏         | 5/250 [00:12<08:54,  2.18s/it]\n",
      "Loss: 0.03486:   2%|▏         | 5/250 [00:12<08:54,  2.18s/it]\n",
      "Loss: 0.03162:   2%|▏         | 5/250 [00:12<08:54,  2.18s/it]\n",
      "Loss: 0.03143:   2%|▏         | 5/250 [00:12<08:54,  2.18s/it]\n",
      "Loss: 0.03294:   2%|▏         | 5/250 [00:12<08:54,  2.18s/it]\n",
      "Loss: 0.03225:   2%|▏         | 5/250 [00:12<08:54,  2.18s/it]\n",
      "Loss: 0.02855:   2%|▏         | 5/250 [00:12<08:54,  2.18s/it]\n",
      "Loss: 0.02855:   2%|▏         | 6/250 [00:12<08:42,  2.14s/it]\n",
      "Loss: 0.03259:   2%|▏         | 6/250 [00:12<08:42,  2.14s/it]\n",
      "Loss: 0.02836:   2%|▏         | 6/250 [00:13<08:42,  2.14s/it]\n",
      "Loss: 0.03199:   2%|▏         | 6/250 [00:13<08:42,  2.14s/it]\n",
      "Loss: 0.03576:   2%|▏         | 6/250 [00:13<08:42,  2.14s/it]\n",
      "Loss: 0.03390:   2%|▏         | 6/250 [00:13<08:42,  2.14s/it]\n",
      "Loss: 0.03323:   2%|▏         | 6/250 [00:13<08:42,  2.14s/it]\n",
      "Loss: 0.03201:   2%|▏         | 6/250 [00:13<08:42,  2.14s/it]\n",
      "Loss: 0.03352:   2%|▏         | 6/250 [00:13<08:42,  2.14s/it]\n",
      "Loss: 0.03259:   2%|▏         | 6/250 [00:13<08:42,  2.14s/it]\n",
      "Loss: 0.03549:   2%|▏         | 6/250 [00:13<08:42,  2.14s/it]\n",
      "Loss: 0.03386:   2%|▏         | 6/250 [00:14<08:42,  2.14s/it]\n",
      "Loss: 0.02685:   2%|▏         | 6/250 [00:14<08:42,  2.14s/it]\n",
      "Loss: 0.03663:   2%|▏         | 6/250 [00:14<08:42,  2.14s/it]\n",
      "Loss: 0.02950:   2%|▏         | 6/250 [00:14<08:42,  2.14s/it]\n",
      "Loss: 0.02812:   2%|▏         | 6/250 [00:14<08:42,  2.14s/it]\n",
      "Loss: 0.03369:   2%|▏         | 6/250 [00:14<08:42,  2.14s/it]\n",
      "Loss: 0.02935:   2%|▏         | 6/250 [00:14<08:42,  2.14s/it]\n",
      "Loss: 0.03151:   2%|▏         | 6/250 [00:14<08:42,  2.14s/it]\n",
      "Loss: 0.02714:   2%|▏         | 6/250 [00:14<08:42,  2.14s/it]\n",
      "Loss: 0.02714:   3%|▎         | 7/250 [00:14<08:34,  2.12s/it]\n",
      "Loss: 0.02990:   3%|▎         | 7/250 [00:14<08:34,  2.12s/it]\n",
      "Loss: 0.03173:   3%|▎         | 7/250 [00:15<08:34,  2.12s/it]\n",
      "Loss: 0.03231:   3%|▎         | 7/250 [00:15<08:34,  2.12s/it]\n",
      "Loss: 0.02949:   3%|▎         | 7/250 [00:15<08:34,  2.12s/it]\n",
      "Loss: 0.02976:   3%|▎         | 7/250 [00:15<08:34,  2.12s/it]\n",
      "Loss: 0.02894:   3%|▎         | 7/250 [00:15<08:34,  2.12s/it]\n",
      "Loss: 0.02823:   3%|▎         | 7/250 [00:15<08:34,  2.12s/it]\n",
      "Loss: 0.03275:   3%|▎         | 7/250 [00:15<08:34,  2.12s/it]\n",
      "Loss: 0.03263:   3%|▎         | 7/250 [00:15<08:34,  2.12s/it]\n",
      "Loss: 0.03176:   3%|▎         | 7/250 [00:15<08:34,  2.12s/it]\n",
      "Loss: 0.03228:   3%|▎         | 7/250 [00:16<08:34,  2.12s/it]\n",
      "Loss: 0.03078:   3%|▎         | 7/250 [00:16<08:34,  2.12s/it]\n",
      "Loss: 0.03068:   3%|▎         | 7/250 [00:16<08:34,  2.12s/it]\n",
      "Loss: 0.03028:   3%|▎         | 7/250 [00:16<08:34,  2.12s/it]\n",
      "Loss: 0.03293:   3%|▎         | 7/250 [00:16<08:34,  2.12s/it]\n",
      "Loss: 0.02726:   3%|▎         | 7/250 [00:16<08:34,  2.12s/it]\n",
      "Loss: 0.03234:   3%|▎         | 7/250 [00:16<08:34,  2.12s/it]\n",
      "Loss: 0.03756:   3%|▎         | 7/250 [00:16<08:34,  2.12s/it]\n",
      "Loss: 0.02741:   3%|▎         | 7/250 [00:16<08:34,  2.12s/it]\n",
      "Loss: 0.02741:   3%|▎         | 8/250 [00:16<08:26,  2.09s/it]\n",
      "Loss: 0.02972:   3%|▎         | 8/250 [00:17<08:26,  2.09s/it]\n",
      "Loss: 0.02516:   3%|▎         | 8/250 [00:17<08:26,  2.09s/it]\n",
      "Loss: 0.03255:   3%|▎         | 8/250 [00:17<08:26,  2.09s/it]\n",
      "Loss: 0.03741:   3%|▎         | 8/250 [00:17<08:26,  2.09s/it]\n",
      "Loss: 0.02662:   3%|▎         | 8/250 [00:17<08:26,  2.09s/it]\n",
      "Loss: 0.02889:   3%|▎         | 8/250 [00:17<08:26,  2.09s/it]\n",
      "Loss: 0.02713:   3%|▎         | 8/250 [00:17<08:26,  2.09s/it]\n",
      "Loss: 0.02847:   3%|▎         | 8/250 [00:17<08:26,  2.09s/it]\n",
      "Loss: 0.03154:   3%|▎         | 8/250 [00:17<08:26,  2.09s/it]\n",
      "Loss: 0.03039:   3%|▎         | 8/250 [00:18<08:26,  2.09s/it]\n",
      "Loss: 0.02564:   3%|▎         | 8/250 [00:18<08:26,  2.09s/it]\n",
      "Loss: 0.02419:   3%|▎         | 8/250 [00:18<08:26,  2.09s/it]\n",
      "Loss: 0.02965:   3%|▎         | 8/250 [00:18<08:26,  2.09s/it]\n",
      "Loss: 0.02658:   3%|▎         | 8/250 [00:18<08:26,  2.09s/it]\n",
      "Loss: 0.02821:   3%|▎         | 8/250 [00:18<08:26,  2.09s/it]\n",
      "Loss: 0.02727:   3%|▎         | 8/250 [00:18<08:26,  2.09s/it]\n",
      "Loss: 0.03062:   3%|▎         | 8/250 [00:18<08:26,  2.09s/it]\n",
      "Loss: 0.02585:   3%|▎         | 8/250 [00:18<08:26,  2.09s/it]\n",
      "Loss: 0.02436:   3%|▎         | 8/250 [00:18<08:26,  2.09s/it]\n",
      "Loss: 0.02436:   4%|▎         | 9/250 [00:18<08:23,  2.09s/it]\n",
      "Loss: 0.02775:   4%|▎         | 9/250 [00:19<08:23,  2.09s/it]\n",
      "Loss: 0.02671:   4%|▎         | 9/250 [00:19<08:23,  2.09s/it]\n",
      "Loss: 0.03575:   4%|▎         | 9/250 [00:19<08:23,  2.09s/it]\n",
      "Loss: 0.02852:   4%|▎         | 9/250 [00:19<08:23,  2.09s/it]\n",
      "Loss: 0.02421:   4%|▎         | 9/250 [00:19<08:23,  2.09s/it]\n",
      "Loss: 0.02805:   4%|▎         | 9/250 [00:19<08:23,  2.09s/it]\n",
      "Loss: 0.02879:   4%|▎         | 9/250 [00:19<08:23,  2.09s/it]\n",
      "Loss: 0.03320:   4%|▎         | 9/250 [00:19<08:23,  2.09s/it]\n",
      "Loss: 0.02791:   4%|▎         | 9/250 [00:19<08:23,  2.09s/it]\n",
      "Loss: 0.02986:   4%|▎         | 9/250 [00:20<08:23,  2.09s/it]\n",
      "Loss: 0.02823:   4%|▎         | 9/250 [00:20<08:23,  2.09s/it]\n",
      "Loss: 0.03253:   4%|▎         | 9/250 [00:20<08:23,  2.09s/it]\n",
      "Loss: 0.02904:   4%|▎         | 9/250 [00:20<08:23,  2.09s/it]\n",
      "Loss: 0.02796:   4%|▎         | 9/250 [00:20<08:23,  2.09s/it]\n",
      "Loss: 0.02528:   4%|▎         | 9/250 [00:20<08:23,  2.09s/it]\n",
      "Loss: 0.02699:   4%|▎         | 9/250 [00:20<08:23,  2.09s/it]\n",
      "Loss: 0.03223:   4%|▎         | 9/250 [00:20<08:23,  2.09s/it]\n",
      "Loss: 0.02714:   4%|▎         | 9/250 [00:20<08:23,  2.09s/it]\n",
      "Loss: 0.02353:   4%|▎         | 9/250 [00:21<08:23,  2.09s/it]\n",
      "Loss: 0.02353:   4%|▍         | 10/250 [00:21<08:20,  2.08s/it]\n",
      "Loss: 0.02680:   4%|▍         | 10/250 [00:21<08:20,  2.08s/it]\n",
      "Loss: 0.02806:   4%|▍         | 10/250 [00:21<08:20,  2.08s/it]\n",
      "Loss: 0.02968:   4%|▍         | 10/250 [00:21<08:20,  2.08s/it]\n",
      "Loss: 0.02573:   4%|▍         | 10/250 [00:21<08:20,  2.08s/it]\n",
      "Loss: 0.02580:   4%|▍         | 10/250 [00:21<08:20,  2.08s/it]\n",
      "Loss: 0.02530:   4%|▍         | 10/250 [00:21<08:20,  2.08s/it]\n",
      "Loss: 0.02691:   4%|▍         | 10/250 [00:21<08:20,  2.08s/it]\n",
      "Loss: 0.02743:   4%|▍         | 10/250 [00:21<08:20,  2.08s/it]\n",
      "Loss: 0.02566:   4%|▍         | 10/250 [00:22<08:20,  2.08s/it]\n",
      "Loss: 0.02994:   4%|▍         | 10/250 [00:22<08:20,  2.08s/it]\n",
      "Loss: 0.03043:   4%|▍         | 10/250 [00:22<08:20,  2.08s/it]\n",
      "Loss: 0.02832:   4%|▍         | 10/250 [00:22<08:20,  2.08s/it]\n",
      "Loss: 0.02735:   4%|▍         | 10/250 [00:22<08:20,  2.08s/it]\n",
      "Loss: 0.02550:   4%|▍         | 10/250 [00:22<08:20,  2.08s/it]\n",
      "Loss: 0.02713:   4%|▍         | 10/250 [00:22<08:20,  2.08s/it]\n",
      "Loss: 0.02897:   4%|▍         | 10/250 [00:22<08:20,  2.08s/it]\n",
      "Loss: 0.02674:   4%|▍         | 10/250 [00:22<08:20,  2.08s/it]\n",
      "Loss: 0.02333:   4%|▍         | 10/250 [00:23<08:20,  2.08s/it]\n",
      "Loss: 0.02333:   4%|▍         | 11/250 [00:23<08:13,  2.07s/it]\n",
      "Loss: 0.02508:   4%|▍         | 11/250 [00:23<08:13,  2.07s/it]\n",
      "Loss: 0.02411:   4%|▍         | 11/250 [00:23<08:13,  2.07s/it]\n",
      "Loss: 0.02887:   4%|▍         | 11/250 [00:23<08:13,  2.07s/it]\n",
      "Loss: 0.02868:   4%|▍         | 11/250 [00:23<08:13,  2.07s/it]\n",
      "Loss: 0.02972:   4%|▍         | 11/250 [00:23<08:13,  2.07s/it]\n",
      "Loss: 0.02883:   4%|▍         | 11/250 [00:23<08:13,  2.07s/it]\n",
      "Loss: 0.02668:   4%|▍         | 11/250 [00:23<08:13,  2.07s/it]\n",
      "Loss: 0.02873:   4%|▍         | 11/250 [00:23<08:13,  2.07s/it]\n",
      "Loss: 0.02864:   4%|▍         | 11/250 [00:24<08:13,  2.07s/it]\n",
      "Loss: 0.02891:   4%|▍         | 11/250 [00:24<08:13,  2.07s/it]\n",
      "Loss: 0.02867:   4%|▍         | 11/250 [00:24<08:13,  2.07s/it]\n",
      "Loss: 0.02729:   4%|▍         | 11/250 [00:24<08:13,  2.07s/it]\n",
      "Loss: 0.02704:   4%|▍         | 11/250 [00:24<08:13,  2.07s/it]\n",
      "Loss: 0.02486:   4%|▍         | 11/250 [00:24<08:13,  2.07s/it]\n",
      "Loss: 0.02637:   4%|▍         | 11/250 [00:24<08:13,  2.07s/it]\n",
      "Loss: 0.02649:   4%|▍         | 11/250 [00:24<08:13,  2.07s/it]\n",
      "Loss: 0.02987:   4%|▍         | 11/250 [00:24<08:13,  2.07s/it]\n",
      "Loss: 0.02656:   4%|▍         | 11/250 [00:25<08:13,  2.07s/it]\n",
      "Loss: 0.02271:   4%|▍         | 11/250 [00:25<08:13,  2.07s/it]\n",
      "Loss: 0.02271:   5%|▍         | 12/250 [00:25<08:11,  2.06s/it]\n",
      "Loss: 0.02749:   5%|▍         | 12/250 [00:25<08:11,  2.06s/it]\n",
      "Loss: 0.02833:   5%|▍         | 12/250 [00:25<08:11,  2.06s/it]\n",
      "Loss: 0.02939:   5%|▍         | 12/250 [00:25<08:11,  2.06s/it]\n",
      "Loss: 0.02699:   5%|▍         | 12/250 [00:25<08:11,  2.06s/it]\n",
      "Loss: 0.02527:   5%|▍         | 12/250 [00:25<08:11,  2.06s/it]\n",
      "Loss: 0.02670:   5%|▍         | 12/250 [00:25<08:11,  2.06s/it]\n",
      "Loss: 0.02635:   5%|▍         | 12/250 [00:25<08:11,  2.06s/it]\n",
      "Loss: 0.03000:   5%|▍         | 12/250 [00:26<08:11,  2.06s/it]\n",
      "Loss: 0.02676:   5%|▍         | 12/250 [00:26<08:11,  2.06s/it]\n",
      "Loss: 0.02641:   5%|▍         | 12/250 [00:26<08:11,  2.06s/it]\n",
      "Loss: 0.02895:   5%|▍         | 12/250 [00:26<08:11,  2.06s/it]\n",
      "Loss: 0.02783:   5%|▍         | 12/250 [00:26<08:11,  2.06s/it]\n",
      "Loss: 0.02745:   5%|▍         | 12/250 [00:26<08:11,  2.06s/it]\n",
      "Loss: 0.02878:   5%|▍         | 12/250 [00:26<08:11,  2.06s/it]\n",
      "Loss: 0.02788:   5%|▍         | 12/250 [00:26<08:11,  2.06s/it]\n",
      "Loss: 0.02644:   5%|▍         | 12/250 [00:26<08:11,  2.06s/it]\n",
      "Loss: 0.02734:   5%|▍         | 12/250 [00:27<08:11,  2.06s/it]\n",
      "Loss: 0.02928:   5%|▍         | 12/250 [00:27<08:11,  2.06s/it]\n",
      "Loss: 0.02339:   5%|▍         | 12/250 [00:27<08:11,  2.06s/it]\n",
      "Loss: 0.02339:   5%|▌         | 13/250 [00:27<08:07,  2.06s/it]\n",
      "Loss: 0.02453:   5%|▌         | 13/250 [00:27<08:07,  2.06s/it]\n",
      "Loss: 0.02297:   5%|▌         | 13/250 [00:27<08:07,  2.06s/it]\n",
      "Loss: 0.02438:   5%|▌         | 13/250 [00:27<08:07,  2.06s/it]\n",
      "Loss: 0.02463:   5%|▌         | 13/250 [00:27<08:07,  2.06s/it]\n",
      "Loss: 0.02925:   5%|▌         | 13/250 [00:27<08:07,  2.06s/it]\n",
      "Loss: 0.02846:   5%|▌         | 13/250 [00:27<08:07,  2.06s/it]\n",
      "Loss: 0.02686:   5%|▌         | 13/250 [00:27<08:07,  2.06s/it]\n",
      "Loss: 0.02747:   5%|▌         | 13/250 [00:28<08:07,  2.06s/it]\n",
      "Loss: 0.02739:   5%|▌         | 13/250 [00:28<08:07,  2.06s/it]\n",
      "Loss: 0.02922:   5%|▌         | 13/250 [00:28<08:07,  2.06s/it]\n",
      "Loss: 0.02856:   5%|▌         | 13/250 [00:28<08:07,  2.06s/it]\n",
      "Loss: 0.02484:   5%|▌         | 13/250 [00:28<08:07,  2.06s/it]\n",
      "Loss: 0.03028:   5%|▌         | 13/250 [00:28<08:07,  2.06s/it]\n",
      "Loss: 0.02596:   5%|▌         | 13/250 [00:28<08:07,  2.06s/it]\n",
      "Loss: 0.02377:   5%|▌         | 13/250 [00:28<08:07,  2.06s/it]\n",
      "Loss: 0.02840:   5%|▌         | 13/250 [00:28<08:07,  2.06s/it]\n",
      "Loss: 0.02419:   5%|▌         | 13/250 [00:29<08:07,  2.06s/it]\n",
      "Loss: 0.02769:   5%|▌         | 13/250 [00:29<08:07,  2.06s/it]\n",
      "Loss: 0.02241:   5%|▌         | 13/250 [00:29<08:07,  2.06s/it]\n",
      "Loss: 0.02241:   6%|▌         | 14/250 [00:29<08:04,  2.05s/it]\n",
      "Loss: 0.02591:   6%|▌         | 14/250 [00:29<08:04,  2.05s/it]\n",
      "Loss: 0.02874:   6%|▌         | 14/250 [00:29<08:04,  2.05s/it]\n",
      "Loss: 0.02943:   6%|▌         | 14/250 [00:29<08:04,  2.05s/it]\n",
      "Loss: 0.02745:   6%|▌         | 14/250 [00:29<08:04,  2.05s/it]\n",
      "Loss: 0.02731:   6%|▌         | 14/250 [00:29<08:04,  2.05s/it]\n",
      "Loss: 0.02315:   6%|▌         | 14/250 [00:29<08:04,  2.05s/it]\n",
      "Loss: 0.02514:   6%|▌         | 14/250 [00:30<08:04,  2.05s/it]\n",
      "Loss: 0.02534:   6%|▌         | 14/250 [00:30<08:04,  2.05s/it]\n",
      "Loss: 0.02831:   6%|▌         | 14/250 [00:30<08:04,  2.05s/it]\n",
      "Loss: 0.02719:   6%|▌         | 14/250 [00:30<08:04,  2.05s/it]\n",
      "Loss: 0.02414:   6%|▌         | 14/250 [00:30<08:04,  2.05s/it]\n",
      "Loss: 0.02558:   6%|▌         | 14/250 [00:30<08:04,  2.05s/it]\n",
      "Loss: 0.02683:   6%|▌         | 14/250 [00:30<08:04,  2.05s/it]\n",
      "Loss: 0.02379:   6%|▌         | 14/250 [00:30<08:04,  2.05s/it]\n",
      "Loss: 0.02942:   6%|▌         | 14/250 [00:30<08:04,  2.05s/it]\n",
      "Loss: 0.03207:   6%|▌         | 14/250 [00:31<08:04,  2.05s/it]\n",
      "Loss: 0.03045:   6%|▌         | 14/250 [00:31<08:04,  2.05s/it]\n",
      "Loss: 0.02533:   6%|▌         | 14/250 [00:31<08:04,  2.05s/it]\n",
      "Loss: 0.02229:   6%|▌         | 14/250 [00:31<08:04,  2.05s/it]\n",
      "Loss: 0.02229:   6%|▌         | 15/250 [00:31<08:02,  2.05s/it]\n",
      "Loss: 0.02601:   6%|▌         | 15/250 [00:31<08:02,  2.05s/it]\n",
      "Loss: 0.02668:   6%|▌         | 15/250 [00:31<08:02,  2.05s/it]\n",
      "Loss: 0.02886:   6%|▌         | 15/250 [00:31<08:02,  2.05s/it]\n",
      "Loss: 0.02635:   6%|▌         | 15/250 [00:31<08:02,  2.05s/it]\n",
      "Loss: 0.02813:   6%|▌         | 15/250 [00:31<08:02,  2.05s/it]\n",
      "Loss: 0.02596:   6%|▌         | 15/250 [00:31<08:02,  2.05s/it]\n",
      "Loss: 0.02498:   6%|▌         | 15/250 [00:32<08:02,  2.05s/it]\n",
      "Loss: 0.02637:   6%|▌         | 15/250 [00:32<08:02,  2.05s/it]\n",
      "Loss: 0.02817:   6%|▌         | 15/250 [00:32<08:02,  2.05s/it]\n",
      "Loss: 0.02681:   6%|▌         | 15/250 [00:32<08:02,  2.05s/it]\n",
      "Loss: 0.02670:   6%|▌         | 15/250 [00:32<08:02,  2.05s/it]\n",
      "Loss: 0.02594:   6%|▌         | 15/250 [00:32<08:02,  2.05s/it]\n",
      "Loss: 0.02714:   6%|▌         | 15/250 [00:32<08:02,  2.05s/it]\n",
      "Loss: 0.02317:   6%|▌         | 15/250 [00:32<08:02,  2.05s/it]\n",
      "Loss: 0.02854:   6%|▌         | 15/250 [00:32<08:02,  2.05s/it]\n",
      "Loss: 0.03169:   6%|▌         | 15/250 [00:33<08:02,  2.05s/it]\n",
      "Loss: 0.03058:   6%|▌         | 15/250 [00:33<08:02,  2.05s/it]\n",
      "Loss: 0.02529:   6%|▌         | 15/250 [00:33<08:02,  2.05s/it]\n",
      "Loss: 0.02312:   6%|▌         | 15/250 [00:33<08:02,  2.05s/it]\n",
      "Loss: 0.02312:   6%|▋         | 16/250 [00:33<07:59,  2.05s/it]\n",
      "Loss: 0.02542:   6%|▋         | 16/250 [00:33<07:59,  2.05s/it]\n",
      "Loss: 0.02670:   6%|▋         | 16/250 [00:33<07:59,  2.05s/it]\n",
      "Loss: 0.02824:   6%|▋         | 16/250 [00:33<07:59,  2.05s/it]\n",
      "Loss: 0.02522:   6%|▋         | 16/250 [00:33<07:59,  2.05s/it]\n",
      "Loss: 0.02609:   6%|▋         | 16/250 [00:33<07:59,  2.05s/it]\n",
      "Loss: 0.02330:   6%|▋         | 16/250 [00:33<07:59,  2.05s/it]\n",
      "Loss: 0.02445:   6%|▋         | 16/250 [00:34<07:59,  2.05s/it]\n",
      "Loss: 0.02914:   6%|▋         | 16/250 [00:34<07:59,  2.05s/it]\n",
      "Loss: 0.02579:   6%|▋         | 16/250 [00:34<07:59,  2.05s/it]\n",
      "Loss: 0.03023:   6%|▋         | 16/250 [00:34<07:59,  2.05s/it]\n",
      "Loss: 0.02548:   6%|▋         | 16/250 [00:34<07:59,  2.05s/it]\n",
      "Loss: 0.02766:   6%|▋         | 16/250 [00:34<07:59,  2.05s/it]\n",
      "Loss: 0.02526:   6%|▋         | 16/250 [00:34<07:59,  2.05s/it]\n",
      "Loss: 0.02659:   6%|▋         | 16/250 [00:34<07:59,  2.05s/it]\n",
      "Loss: 0.02369:   6%|▋         | 16/250 [00:35<07:59,  2.05s/it]\n",
      "Loss: 0.02563:   6%|▋         | 16/250 [00:35<07:59,  2.05s/it]\n",
      "Loss: 0.02556:   6%|▋         | 16/250 [00:35<07:59,  2.05s/it]\n",
      "Loss: 0.02136:   6%|▋         | 16/250 [00:35<07:59,  2.05s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 62.14it/s]\u001b[A\n",
      "Loss: 0.02136:   7%|▋         | 17/250 [00:35<07:56,  2.04s/it]\n",
      "Loss: 0.02592:   7%|▋         | 17/250 [00:35<07:56,  2.04s/it]\n",
      "Loss: 0.02659:   7%|▋         | 17/250 [00:35<07:56,  2.04s/it]\n",
      "Loss: 0.02888:   7%|▋         | 17/250 [00:35<07:56,  2.04s/it]\n",
      "Loss: 0.02568:   7%|▋         | 17/250 [00:35<07:56,  2.04s/it]\n",
      "Loss: 0.02623:   7%|▋         | 17/250 [00:35<07:56,  2.04s/it]\n",
      "Loss: 0.02319:   7%|▋         | 17/250 [00:36<07:56,  2.04s/it]\n",
      "Loss: 0.02439:   7%|▋         | 17/250 [00:36<07:56,  2.04s/it]\n",
      "Loss: 0.02756:   7%|▋         | 17/250 [00:36<07:56,  2.04s/it]\n",
      "Loss: 0.02526:   7%|▋         | 17/250 [00:36<07:56,  2.04s/it]\n",
      "Loss: 0.02818:   7%|▋         | 17/250 [00:36<07:56,  2.04s/it]\n",
      "Loss: 0.02580:   7%|▋         | 17/250 [00:36<07:56,  2.04s/it]\n",
      "Loss: 0.02786:   7%|▋         | 17/250 [00:36<07:56,  2.04s/it]\n",
      "Loss: 0.02568:   7%|▋         | 17/250 [00:36<07:56,  2.04s/it]\n",
      "Loss: 0.02616:   7%|▋         | 17/250 [00:36<07:56,  2.04s/it]\n",
      "Loss: 0.02369:   7%|▋         | 17/250 [00:37<07:56,  2.04s/it]\n",
      "Loss: 0.02488:   7%|▋         | 17/250 [00:37<07:56,  2.04s/it]\n",
      "Loss: 0.02583:   7%|▋         | 17/250 [00:37<07:56,  2.04s/it]\n",
      "Loss: 0.02206:   7%|▋         | 17/250 [00:37<07:56,  2.04s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.96it/s]\u001b[A\n",
      "Loss: 0.02206:   7%|▋         | 18/250 [00:37<07:52,  2.03s/it]\n",
      "Loss: 0.02586:   7%|▋         | 18/250 [00:37<07:52,  2.03s/it]\n",
      "Loss: 0.02622:   7%|▋         | 18/250 [00:37<07:52,  2.03s/it]\n",
      "Loss: 0.02902:   7%|▋         | 18/250 [00:37<07:52,  2.03s/it]\n",
      "Loss: 0.02505:   7%|▋         | 18/250 [00:37<07:52,  2.03s/it]\n",
      "Loss: 0.02410:   7%|▋         | 18/250 [00:37<07:52,  2.03s/it]\n",
      "Loss: 0.02468:   7%|▋         | 18/250 [00:38<07:52,  2.03s/it]\n",
      "Loss: 0.02539:   7%|▋         | 18/250 [00:38<07:52,  2.03s/it]\n",
      "Loss: 0.02868:   7%|▋         | 18/250 [00:38<07:52,  2.03s/it]\n",
      "Loss: 0.02497:   7%|▋         | 18/250 [00:38<07:52,  2.03s/it]\n",
      "Loss: 0.02598:   7%|▋         | 18/250 [00:38<07:52,  2.03s/it]\n",
      "Loss: 0.03052:   7%|▋         | 18/250 [00:38<07:52,  2.03s/it]\n",
      "Loss: 0.02695:   7%|▋         | 18/250 [00:38<07:52,  2.03s/it]\n",
      "Loss: 0.02641:   7%|▋         | 18/250 [00:38<07:52,  2.03s/it]\n",
      "Loss: 0.02285:   7%|▋         | 18/250 [00:38<07:52,  2.03s/it]\n",
      "Loss: 0.02707:   7%|▋         | 18/250 [00:39<07:52,  2.03s/it]\n",
      "Loss: 0.03272:   7%|▋         | 18/250 [00:39<07:52,  2.03s/it]\n",
      "Loss: 0.03018:   7%|▋         | 18/250 [00:39<07:52,  2.03s/it]\n",
      "Loss: 0.02465:   7%|▋         | 18/250 [00:39<07:52,  2.03s/it]\n",
      "Loss: 0.02152:   7%|▋         | 18/250 [00:39<07:52,  2.03s/it]\n",
      "Loss: 0.02152:   8%|▊         | 19/250 [00:39<07:50,  2.04s/it]\n",
      "Loss: 0.02552:   8%|▊         | 19/250 [00:39<07:50,  2.04s/it]\n",
      "Loss: 0.02625:   8%|▊         | 19/250 [00:39<07:50,  2.04s/it]\n",
      "Loss: 0.02830:   8%|▊         | 19/250 [00:39<07:50,  2.04s/it]\n",
      "Loss: 0.02463:   8%|▊         | 19/250 [00:39<07:50,  2.04s/it]\n",
      "Loss: 0.02705:   8%|▊         | 19/250 [00:39<07:50,  2.04s/it]\n",
      "Loss: 0.02273:   8%|▊         | 19/250 [00:40<07:50,  2.04s/it]\n",
      "Loss: 0.02690:   8%|▊         | 19/250 [00:40<07:50,  2.04s/it]\n",
      "Loss: 0.02802:   8%|▊         | 19/250 [00:40<07:50,  2.04s/it]\n",
      "Loss: 0.02521:   8%|▊         | 19/250 [00:40<07:50,  2.04s/it]\n",
      "Loss: 0.02828:   8%|▊         | 19/250 [00:40<07:50,  2.04s/it]\n",
      "Loss: 0.02505:   8%|▊         | 19/250 [00:40<07:50,  2.04s/it]\n",
      "Loss: 0.02829:   8%|▊         | 19/250 [00:40<07:50,  2.04s/it]\n",
      "Loss: 0.02452:   8%|▊         | 19/250 [00:40<07:50,  2.04s/it]\n",
      "Loss: 0.02555:   8%|▊         | 19/250 [00:40<07:50,  2.04s/it]\n",
      "Loss: 0.02359:   8%|▊         | 19/250 [00:41<07:50,  2.04s/it]\n",
      "Loss: 0.02389:   8%|▊         | 19/250 [00:41<07:50,  2.04s/it]\n",
      "Loss: 0.02557:   8%|▊         | 19/250 [00:41<07:50,  2.04s/it]\n",
      "Loss: 0.02122:   8%|▊         | 19/250 [00:41<07:50,  2.04s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 60.24it/s]\u001b[A\n",
      "Loss: 0.02122:   8%|▊         | 20/250 [00:41<07:48,  2.04s/it]\n",
      "Loss: 0.02423:   8%|▊         | 20/250 [00:41<07:48,  2.04s/it]\n",
      "Loss: 0.02425:   8%|▊         | 20/250 [00:41<07:48,  2.04s/it]\n",
      "Loss: 0.02463:   8%|▊         | 20/250 [00:41<07:48,  2.04s/it]\n",
      "Loss: 0.02784:   8%|▊         | 20/250 [00:41<07:48,  2.04s/it]\n",
      "Loss: 0.02126:   8%|▊         | 20/250 [00:41<07:48,  2.04s/it]\n",
      "Loss: 0.02714:   8%|▊         | 20/250 [00:42<07:48,  2.04s/it]\n",
      "Loss: 0.02739:   8%|▊         | 20/250 [00:42<07:48,  2.04s/it]\n",
      "Loss: 0.02546:   8%|▊         | 20/250 [00:42<07:48,  2.04s/it]\n",
      "Loss: 0.02752:   8%|▊         | 20/250 [00:42<07:48,  2.04s/it]\n",
      "Loss: 0.02548:   8%|▊         | 20/250 [00:42<07:48,  2.04s/it]\n",
      "Loss: 0.02421:   8%|▊         | 20/250 [00:42<07:48,  2.04s/it]\n",
      "Loss: 0.02788:   8%|▊         | 20/250 [00:42<07:48,  2.04s/it]\n",
      "Loss: 0.02581:   8%|▊         | 20/250 [00:42<07:48,  2.04s/it]\n",
      "Loss: 0.02761:   8%|▊         | 20/250 [00:42<07:48,  2.04s/it]\n",
      "Loss: 0.02342:   8%|▊         | 20/250 [00:43<07:48,  2.04s/it]\n",
      "Loss: 0.02473:   8%|▊         | 20/250 [00:43<07:48,  2.04s/it]\n",
      "Loss: 0.02554:   8%|▊         | 20/250 [00:43<07:48,  2.04s/it]\n",
      "Loss: 0.02615:   8%|▊         | 20/250 [00:43<07:48,  2.04s/it]\n",
      "Loss: 0.02097:   8%|▊         | 20/250 [00:43<07:48,  2.04s/it]\n",
      "Loss: 0.02097:   8%|▊         | 21/250 [00:43<07:48,  2.04s/it]\n",
      "Loss: 0.02752:   8%|▊         | 21/250 [00:43<07:48,  2.04s/it]\n",
      "Loss: 0.02539:   8%|▊         | 21/250 [00:43<07:48,  2.04s/it]\n",
      "Loss: 0.02767:   8%|▊         | 21/250 [00:43<07:48,  2.04s/it]\n",
      "Loss: 0.02706:   8%|▊         | 21/250 [00:43<07:48,  2.04s/it]\n",
      "Loss: 0.02752:   8%|▊         | 21/250 [00:44<07:48,  2.04s/it]\n",
      "Loss: 0.02374:   8%|▊         | 21/250 [00:44<07:48,  2.04s/it]\n",
      "Loss: 0.03167:   8%|▊         | 21/250 [00:44<07:48,  2.04s/it]\n",
      "Loss: 0.02648:   8%|▊         | 21/250 [00:44<07:48,  2.04s/it]\n",
      "Loss: 0.02854:   8%|▊         | 21/250 [00:44<07:48,  2.04s/it]\n",
      "Loss: 0.02405:   8%|▊         | 21/250 [00:44<07:48,  2.04s/it]\n",
      "Loss: 0.02823:   8%|▊         | 21/250 [00:44<07:48,  2.04s/it]\n",
      "Loss: 0.02830:   8%|▊         | 21/250 [00:44<07:48,  2.04s/it]\n",
      "Loss: 0.02737:   8%|▊         | 21/250 [00:44<07:48,  2.04s/it]\n",
      "Loss: 0.02371:   8%|▊         | 21/250 [00:45<07:48,  2.04s/it]\n",
      "Loss: 0.02483:   8%|▊         | 21/250 [00:45<07:48,  2.04s/it]\n",
      "Loss: 0.02641:   8%|▊         | 21/250 [00:45<07:48,  2.04s/it]\n",
      "Loss: 0.02457:   8%|▊         | 21/250 [00:45<07:48,  2.04s/it]\n",
      "Loss: 0.02768:   8%|▊         | 21/250 [00:45<07:48,  2.04s/it]\n",
      "Loss: 0.02631:   8%|▊         | 21/250 [00:45<07:48,  2.04s/it]\n",
      "Loss: 0.02169:   8%|▊         | 21/250 [00:45<07:48,  2.04s/it]\n",
      "Loss: 0.02169:   9%|▉         | 22/250 [00:45<07:50,  2.06s/it]\n",
      "Loss: 0.02477:   9%|▉         | 22/250 [00:45<07:50,  2.06s/it]\n",
      "Loss: 0.02370:   9%|▉         | 22/250 [00:45<07:50,  2.06s/it]\n",
      "Loss: 0.02423:   9%|▉         | 22/250 [00:45<07:50,  2.06s/it]\n",
      "Loss: 0.02375:   9%|▉         | 22/250 [00:46<07:50,  2.06s/it]\n",
      "Loss: 0.02546:   9%|▉         | 22/250 [00:46<07:50,  2.06s/it]\n",
      "Loss: 0.02471:   9%|▉         | 22/250 [00:46<07:50,  2.06s/it]\n",
      "Loss: 0.02888:   9%|▉         | 22/250 [00:46<07:50,  2.06s/it]\n",
      "Loss: 0.02721:   9%|▉         | 22/250 [00:46<07:50,  2.06s/it]\n",
      "Loss: 0.02684:   9%|▉         | 22/250 [00:46<07:50,  2.06s/it]\n",
      "Loss: 0.02433:   9%|▉         | 22/250 [00:46<07:50,  2.06s/it]\n",
      "Loss: 0.02540:   9%|▉         | 22/250 [00:46<07:50,  2.06s/it]\n",
      "Loss: 0.02313:   9%|▉         | 22/250 [00:46<07:50,  2.06s/it]\n",
      "Loss: 0.02497:   9%|▉         | 22/250 [00:46<07:50,  2.06s/it]\n",
      "Loss: 0.02757:   9%|▉         | 22/250 [00:47<07:50,  2.06s/it]\n",
      "Loss: 0.02477:   9%|▉         | 22/250 [00:47<07:50,  2.06s/it]\n",
      "Loss: 0.02348:   9%|▉         | 22/250 [00:47<07:50,  2.06s/it]\n",
      "Loss: 0.02470:   9%|▉         | 22/250 [00:47<07:50,  2.06s/it]\n",
      "Loss: 0.02718:   9%|▉         | 22/250 [00:47<07:50,  2.06s/it]\n",
      "Loss: 0.02416:   9%|▉         | 22/250 [00:47<07:50,  2.06s/it]\n",
      "Loss: 0.02281:   9%|▉         | 22/250 [00:47<07:50,  2.06s/it]\n",
      "Loss: 0.02281:   9%|▉         | 23/250 [00:47<07:51,  2.08s/it]\n",
      "Loss: 0.02503:   9%|▉         | 23/250 [00:47<07:51,  2.08s/it]\n",
      "Loss: 0.02481:   9%|▉         | 23/250 [00:47<07:51,  2.08s/it]\n",
      "Loss: 0.02678:   9%|▉         | 23/250 [00:48<07:51,  2.08s/it]\n",
      "Loss: 0.02393:   9%|▉         | 23/250 [00:48<07:51,  2.08s/it]\n",
      "Loss: 0.02075:   9%|▉         | 23/250 [00:48<07:51,  2.08s/it]\n",
      "Loss: 0.02613:   9%|▉         | 23/250 [00:48<07:51,  2.08s/it]\n",
      "Loss: 0.02417:   9%|▉         | 23/250 [00:48<07:51,  2.08s/it]\n",
      "Loss: 0.02811:   9%|▉         | 23/250 [00:48<07:51,  2.08s/it]\n",
      "Loss: 0.02586:   9%|▉         | 23/250 [00:48<07:51,  2.08s/it]\n",
      "Loss: 0.02315:   9%|▉         | 23/250 [00:48<07:51,  2.08s/it]\n",
      "Loss: 0.02255:   9%|▉         | 23/250 [00:48<07:51,  2.08s/it]\n",
      "Loss: 0.02506:   9%|▉         | 23/250 [00:49<07:51,  2.08s/it]\n",
      "Loss: 0.02503:   9%|▉         | 23/250 [00:49<07:51,  2.08s/it]\n",
      "Loss: 0.02553:   9%|▉         | 23/250 [00:49<07:51,  2.08s/it]\n",
      "Loss: 0.02358:   9%|▉         | 23/250 [00:49<07:51,  2.08s/it]\n",
      "Loss: 0.02419:   9%|▉         | 23/250 [00:49<07:51,  2.08s/it]\n",
      "Loss: 0.02748:   9%|▉         | 23/250 [00:49<07:51,  2.08s/it]\n",
      "Loss: 0.02399:   9%|▉         | 23/250 [00:49<07:51,  2.08s/it]\n",
      "Loss: 0.02036:   9%|▉         | 23/250 [00:49<07:51,  2.08s/it]\n",
      "Loss: 0.02036:  10%|▉         | 24/250 [00:49<07:49,  2.08s/it]\n",
      "Loss: 0.02545:  10%|▉         | 24/250 [00:49<07:49,  2.08s/it]\n",
      "Loss: 0.02473:  10%|▉         | 24/250 [00:50<07:49,  2.08s/it]\n",
      "Loss: 0.02710:  10%|▉         | 24/250 [00:50<07:49,  2.08s/it]\n",
      "Loss: 0.02434:  10%|▉         | 24/250 [00:50<07:49,  2.08s/it]\n",
      "Loss: 0.02121:  10%|▉         | 24/250 [00:50<07:49,  2.08s/it]\n",
      "Loss: 0.02467:  10%|▉         | 24/250 [00:50<07:49,  2.08s/it]\n",
      "Loss: 0.02277:  10%|▉         | 24/250 [00:50<07:49,  2.08s/it]\n",
      "Loss: 0.02617:  10%|▉         | 24/250 [00:50<07:49,  2.08s/it]\n",
      "Loss: 0.02649:  10%|▉         | 24/250 [00:50<07:49,  2.08s/it]\n",
      "Loss: 0.02476:  10%|▉         | 24/250 [00:50<07:49,  2.08s/it]\n",
      "Loss: 0.02570:  10%|▉         | 24/250 [00:50<07:49,  2.08s/it]\n",
      "Loss: 0.02520:  10%|▉         | 24/250 [00:51<07:49,  2.08s/it]\n",
      "Loss: 0.02386:  10%|▉         | 24/250 [00:51<07:49,  2.08s/it]\n",
      "Loss: 0.02280:  10%|▉         | 24/250 [00:51<07:49,  2.08s/it]\n",
      "Loss: 0.02238:  10%|▉         | 24/250 [00:51<07:49,  2.08s/it]\n",
      "Loss: 0.02348:  10%|▉         | 24/250 [00:51<07:49,  2.08s/it]\n",
      "Loss: 0.02545:  10%|▉         | 24/250 [00:51<07:49,  2.08s/it]\n",
      "Loss: 0.02375:  10%|▉         | 24/250 [00:51<07:49,  2.08s/it]\n",
      "Loss: 0.02071:  10%|▉         | 24/250 [00:51<07:49,  2.08s/it]\n",
      "Loss: 0.02071:  10%|█         | 25/250 [00:51<07:46,  2.08s/it]\n",
      "Loss: 0.02405:  10%|█         | 25/250 [00:51<07:46,  2.08s/it]\n",
      "Loss: 0.02351:  10%|█         | 25/250 [00:52<07:46,  2.08s/it]\n",
      "Loss: 0.02761:  10%|█         | 25/250 [00:52<07:46,  2.08s/it]\n",
      "Loss: 0.02219:  10%|█         | 25/250 [00:52<07:46,  2.08s/it]\n",
      "Loss: 0.02573:  10%|█         | 25/250 [00:52<07:46,  2.08s/it]\n",
      "Loss: 0.02435:  10%|█         | 25/250 [00:52<07:46,  2.08s/it]\n",
      "Loss: 0.02241:  10%|█         | 25/250 [00:52<07:46,  2.08s/it]\n",
      "Loss: 0.02509:  10%|█         | 25/250 [00:52<07:46,  2.08s/it]\n",
      "Loss: 0.02519:  10%|█         | 25/250 [00:52<07:46,  2.08s/it]\n",
      "Loss: 0.02307:  10%|█         | 25/250 [00:52<07:46,  2.08s/it]\n",
      "Loss: 0.02223:  10%|█         | 25/250 [00:53<07:46,  2.08s/it]\n",
      "Loss: 0.02677:  10%|█         | 25/250 [00:53<07:46,  2.08s/it]\n",
      "Loss: 0.02403:  10%|█         | 25/250 [00:53<07:46,  2.08s/it]\n",
      "Loss: 0.02630:  10%|█         | 25/250 [00:53<07:46,  2.08s/it]\n",
      "Loss: 0.02207:  10%|█         | 25/250 [00:53<07:46,  2.08s/it]\n",
      "Loss: 0.02200:  10%|█         | 25/250 [00:53<07:46,  2.08s/it]\n",
      "Loss: 0.02272:  10%|█         | 25/250 [00:53<07:46,  2.08s/it]\n",
      "Loss: 0.02409:  10%|█         | 25/250 [00:53<07:46,  2.08s/it]\n",
      "Loss: 0.01982:  10%|█         | 25/250 [00:53<07:46,  2.08s/it]\n",
      "Loss: 0.01982:  10%|█         | 26/250 [00:53<07:42,  2.07s/it]\n",
      "Loss: 0.02322:  10%|█         | 26/250 [00:54<07:42,  2.07s/it]\n",
      "Loss: 0.02033:  10%|█         | 26/250 [00:54<07:42,  2.07s/it]\n",
      "Loss: 0.02532:  10%|█         | 26/250 [00:54<07:42,  2.07s/it]\n",
      "Loss: 0.02226:  10%|█         | 26/250 [00:54<07:42,  2.07s/it]\n",
      "Loss: 0.02037:  10%|█         | 26/250 [00:54<07:42,  2.07s/it]\n",
      "Loss: 0.02541:  10%|█         | 26/250 [00:54<07:42,  2.07s/it]\n",
      "Loss: 0.02505:  10%|█         | 26/250 [00:54<07:42,  2.07s/it]\n",
      "Loss: 0.02457:  10%|█         | 26/250 [00:54<07:42,  2.07s/it]\n",
      "Loss: 0.02492:  10%|█         | 26/250 [00:54<07:42,  2.07s/it]\n",
      "Loss: 0.02373:  10%|█         | 26/250 [00:55<07:42,  2.07s/it]\n",
      "Loss: 0.02214:  10%|█         | 26/250 [00:55<07:42,  2.07s/it]\n",
      "Loss: 0.02328:  10%|█         | 26/250 [00:55<07:42,  2.07s/it]\n",
      "Loss: 0.02656:  10%|█         | 26/250 [00:55<07:42,  2.07s/it]\n",
      "Loss: 0.02134:  10%|█         | 26/250 [00:55<07:42,  2.07s/it]\n",
      "Loss: 0.02554:  10%|█         | 26/250 [00:55<07:42,  2.07s/it]\n",
      "Loss: 0.02716:  10%|█         | 26/250 [00:55<07:42,  2.07s/it]\n",
      "Loss: 0.02737:  10%|█         | 26/250 [00:55<07:42,  2.07s/it]\n",
      "Loss: 0.02247:  10%|█         | 26/250 [00:55<07:42,  2.07s/it]\n",
      "Loss: 0.01925:  10%|█         | 26/250 [00:55<07:42,  2.07s/it]\n",
      "Loss: 0.01925:  11%|█         | 27/250 [00:55<07:40,  2.07s/it]\n",
      "Loss: 0.02397:  11%|█         | 27/250 [00:56<07:40,  2.07s/it]\n",
      "Loss: 0.02338:  11%|█         | 27/250 [00:56<07:40,  2.07s/it]\n",
      "Loss: 0.02742:  11%|█         | 27/250 [00:56<07:40,  2.07s/it]\n",
      "Loss: 0.02268:  11%|█         | 27/250 [00:56<07:40,  2.07s/it]\n",
      "Loss: 0.02326:  11%|█         | 27/250 [00:56<07:40,  2.07s/it]\n",
      "Loss: 0.02818:  11%|█         | 27/250 [00:56<07:40,  2.07s/it]\n",
      "Loss: 0.02203:  11%|█         | 27/250 [00:56<07:40,  2.07s/it]\n",
      "Loss: 0.02425:  11%|█         | 27/250 [00:56<07:40,  2.07s/it]\n",
      "Loss: 0.02505:  11%|█         | 27/250 [00:56<07:40,  2.07s/it]\n",
      "Loss: 0.02546:  11%|█         | 27/250 [00:57<07:40,  2.07s/it]\n",
      "Loss: 0.02326:  11%|█         | 27/250 [00:57<07:40,  2.07s/it]\n",
      "Loss: 0.02461:  11%|█         | 27/250 [00:57<07:40,  2.07s/it]\n",
      "Loss: 0.02381:  11%|█         | 27/250 [00:57<07:40,  2.07s/it]\n",
      "Loss: 0.02184:  11%|█         | 27/250 [00:57<07:40,  2.07s/it]\n",
      "Loss: 0.02198:  11%|█         | 27/250 [00:57<07:40,  2.07s/it]\n",
      "Loss: 0.02243:  11%|█         | 27/250 [00:57<07:40,  2.07s/it]\n",
      "Loss: 0.02601:  11%|█         | 27/250 [00:57<07:40,  2.07s/it]\n",
      "Loss: 0.02229:  11%|█         | 27/250 [00:57<07:40,  2.07s/it]\n",
      "Loss: 0.02210:  11%|█         | 27/250 [00:58<07:40,  2.07s/it]\n",
      "Loss: 0.02210:  11%|█         | 28/250 [00:58<07:39,  2.07s/it]\n",
      "Loss: 0.02421:  11%|█         | 28/250 [00:58<07:39,  2.07s/it]\n",
      "Loss: 0.02333:  11%|█         | 28/250 [00:58<07:39,  2.07s/it]\n",
      "Loss: 0.02610:  11%|█         | 28/250 [00:58<07:39,  2.07s/it]\n",
      "Loss: 0.02289:  11%|█         | 28/250 [00:58<07:39,  2.07s/it]\n",
      "Loss: 0.02265:  11%|█         | 28/250 [00:58<07:39,  2.07s/it]\n",
      "Loss: 0.02174:  11%|█         | 28/250 [00:58<07:39,  2.07s/it]\n",
      "Loss: 0.02402:  11%|█         | 28/250 [00:58<07:39,  2.07s/it]\n",
      "Loss: 0.02565:  11%|█         | 28/250 [00:58<07:39,  2.07s/it]\n",
      "Loss: 0.02411:  11%|█         | 28/250 [00:59<07:39,  2.07s/it]\n",
      "Loss: 0.02296:  11%|█         | 28/250 [00:59<07:39,  2.07s/it]\n",
      "Loss: 0.02686:  11%|█         | 28/250 [00:59<07:39,  2.07s/it]\n",
      "Loss: 0.02352:  11%|█         | 28/250 [00:59<07:39,  2.07s/it]\n",
      "Loss: 0.02312:  11%|█         | 28/250 [00:59<07:39,  2.07s/it]\n",
      "Loss: 0.02561:  11%|█         | 28/250 [00:59<07:39,  2.07s/it]\n",
      "Loss: 0.02405:  11%|█         | 28/250 [00:59<07:39,  2.07s/it]\n",
      "Loss: 0.02078:  11%|█         | 28/250 [00:59<07:39,  2.07s/it]\n",
      "Loss: 0.02599:  11%|█         | 28/250 [00:59<07:39,  2.07s/it]\n",
      "Loss: 0.02355:  11%|█         | 28/250 [01:00<07:39,  2.07s/it]\n",
      "Loss: 0.01938:  11%|█         | 28/250 [01:00<07:39,  2.07s/it]\n",
      "Loss: 0.01938:  12%|█▏        | 29/250 [01:00<07:33,  2.05s/it]\n",
      "Loss: 0.02396:  12%|█▏        | 29/250 [01:00<07:33,  2.05s/it]\n",
      "Loss: 0.02265:  12%|█▏        | 29/250 [01:00<07:33,  2.05s/it]\n",
      "Loss: 0.02484:  12%|█▏        | 29/250 [01:00<07:33,  2.05s/it]\n",
      "Loss: 0.02111:  12%|█▏        | 29/250 [01:00<07:33,  2.05s/it]\n",
      "Loss: 0.02299:  12%|█▏        | 29/250 [01:00<07:33,  2.05s/it]\n",
      "Loss: 0.02036:  12%|█▏        | 29/250 [01:00<07:33,  2.05s/it]\n",
      "Loss: 0.02232:  12%|█▏        | 29/250 [01:00<07:33,  2.05s/it]\n",
      "Loss: 0.02571:  12%|█▏        | 29/250 [01:00<07:33,  2.05s/it]\n",
      "Loss: 0.02407:  12%|█▏        | 29/250 [01:01<07:33,  2.05s/it]\n",
      "Loss: 0.02255:  12%|█▏        | 29/250 [01:01<07:33,  2.05s/it]\n",
      "Loss: 0.02609:  12%|█▏        | 29/250 [01:01<07:33,  2.05s/it]\n",
      "Loss: 0.02316:  12%|█▏        | 29/250 [01:01<07:33,  2.05s/it]\n",
      "Loss: 0.02383:  12%|█▏        | 29/250 [01:01<07:33,  2.05s/it]\n",
      "Loss: 0.02622:  12%|█▏        | 29/250 [01:01<07:33,  2.05s/it]\n",
      "Loss: 0.02361:  12%|█▏        | 29/250 [01:01<07:33,  2.05s/it]\n",
      "Loss: 0.02134:  12%|█▏        | 29/250 [01:01<07:33,  2.05s/it]\n",
      "Loss: 0.02591:  12%|█▏        | 29/250 [01:01<07:33,  2.05s/it]\n",
      "Loss: 0.02471:  12%|█▏        | 29/250 [01:02<07:33,  2.05s/it]\n",
      "Loss: 0.01955:  12%|█▏        | 29/250 [01:02<07:33,  2.05s/it]\n",
      "Loss: 0.01955:  12%|█▏        | 30/250 [01:02<07:28,  2.04s/it]\n",
      "Loss: 0.02408:  12%|█▏        | 30/250 [01:02<07:28,  2.04s/it]\n",
      "Loss: 0.02190:  12%|█▏        | 30/250 [01:02<07:28,  2.04s/it]\n",
      "Loss: 0.02551:  12%|█▏        | 30/250 [01:02<07:28,  2.04s/it]\n",
      "Loss: 0.02120:  12%|█▏        | 30/250 [01:02<07:28,  2.04s/it]\n",
      "Loss: 0.01933:  12%|█▏        | 30/250 [01:02<07:28,  2.04s/it]\n",
      "Loss: 0.02405:  12%|█▏        | 30/250 [01:02<07:28,  2.04s/it]\n",
      "Loss: 0.02163:  12%|█▏        | 30/250 [01:02<07:28,  2.04s/it]\n",
      "Loss: 0.02320:  12%|█▏        | 30/250 [01:02<07:28,  2.04s/it]\n",
      "Loss: 0.02361:  12%|█▏        | 30/250 [01:03<07:28,  2.04s/it]\n",
      "Loss: 0.02329:  12%|█▏        | 30/250 [01:03<07:28,  2.04s/it]\n",
      "Loss: 0.02158:  12%|█▏        | 30/250 [01:03<07:28,  2.04s/it]\n",
      "Loss: 0.02158:  12%|█▏        | 30/250 [01:03<07:28,  2.04s/it]\n",
      "Loss: 0.02353:  12%|█▏        | 30/250 [01:03<07:28,  2.04s/it]\n",
      "Loss: 0.02166:  12%|█▏        | 30/250 [01:03<07:28,  2.04s/it]\n",
      "Loss: 0.02480:  12%|█▏        | 30/250 [01:03<07:28,  2.04s/it]\n",
      "Loss: 0.02653:  12%|█▏        | 30/250 [01:03<07:28,  2.04s/it]\n",
      "Loss: 0.02601:  12%|█▏        | 30/250 [01:03<07:28,  2.04s/it]\n",
      "Loss: 0.02166:  12%|█▏        | 30/250 [01:04<07:28,  2.04s/it]\n",
      "Loss: 0.01924:  12%|█▏        | 30/250 [01:04<07:28,  2.04s/it]\n",
      "Loss: 0.01924:  12%|█▏        | 31/250 [01:04<07:26,  2.04s/it]\n",
      "Loss: 0.02391:  12%|█▏        | 31/250 [01:04<07:26,  2.04s/it]\n",
      "Loss: 0.02214:  12%|█▏        | 31/250 [01:04<07:26,  2.04s/it]\n",
      "Loss: 0.02494:  12%|█▏        | 31/250 [01:04<07:26,  2.04s/it]\n",
      "Loss: 0.02190:  12%|█▏        | 31/250 [01:04<07:26,  2.04s/it]\n",
      "Loss: 0.02220:  12%|█▏        | 31/250 [01:04<07:26,  2.04s/it]\n",
      "Loss: 0.02189:  12%|█▏        | 31/250 [01:04<07:26,  2.04s/it]\n",
      "Loss: 0.02249:  12%|█▏        | 31/250 [01:04<07:26,  2.04s/it]\n",
      "Loss: 0.02514:  12%|█▏        | 31/250 [01:05<07:26,  2.04s/it]\n",
      "Loss: 0.02277:  12%|█▏        | 31/250 [01:05<07:26,  2.04s/it]\n",
      "Loss: 0.02250:  12%|█▏        | 31/250 [01:05<07:26,  2.04s/it]\n",
      "Loss: 0.02519:  12%|█▏        | 31/250 [01:05<07:26,  2.04s/it]\n",
      "Loss: 0.02222:  12%|█▏        | 31/250 [01:05<07:26,  2.04s/it]\n",
      "Loss: 0.02524:  12%|█▏        | 31/250 [01:05<07:26,  2.04s/it]\n",
      "Loss: 0.01959:  12%|█▏        | 31/250 [01:05<07:26,  2.04s/it]\n",
      "Loss: 0.02481:  12%|█▏        | 31/250 [01:05<07:26,  2.04s/it]\n",
      "Loss: 0.02634:  12%|█▏        | 31/250 [01:05<07:26,  2.04s/it]\n",
      "Loss: 0.02726:  12%|█▏        | 31/250 [01:06<07:26,  2.04s/it]\n",
      "Loss: 0.02081:  12%|█▏        | 31/250 [01:06<07:26,  2.04s/it]\n",
      "Loss: 0.01905:  12%|█▏        | 31/250 [01:06<07:26,  2.04s/it]\n",
      "Loss: 0.01905:  13%|█▎        | 32/250 [01:06<07:26,  2.05s/it]\n",
      "Loss: 0.02278:  13%|█▎        | 32/250 [01:06<07:26,  2.05s/it]\n",
      "Loss: 0.02140:  13%|█▎        | 32/250 [01:06<07:26,  2.05s/it]\n",
      "Loss: 0.02574:  13%|█▎        | 32/250 [01:06<07:26,  2.05s/it]\n",
      "Loss: 0.02103:  13%|█▎        | 32/250 [01:06<07:26,  2.05s/it]\n",
      "Loss: 0.01930:  13%|█▎        | 32/250 [01:06<07:26,  2.05s/it]\n",
      "Loss: 0.02298:  13%|█▎        | 32/250 [01:06<07:26,  2.05s/it]\n",
      "Loss: 0.02163:  13%|█▎        | 32/250 [01:06<07:26,  2.05s/it]\n",
      "Loss: 0.02185:  13%|█▎        | 32/250 [01:07<07:26,  2.05s/it]\n",
      "Loss: 0.02374:  13%|█▎        | 32/250 [01:07<07:26,  2.05s/it]\n",
      "Loss: 0.02368:  13%|█▎        | 32/250 [01:07<07:26,  2.05s/it]\n",
      "Loss: 0.02086:  13%|█▎        | 32/250 [01:07<07:26,  2.05s/it]\n",
      "Loss: 0.02231:  13%|█▎        | 32/250 [01:07<07:26,  2.05s/it]\n",
      "Loss: 0.02275:  13%|█▎        | 32/250 [01:07<07:26,  2.05s/it]\n",
      "Loss: 0.01978:  13%|█▎        | 32/250 [01:07<07:26,  2.05s/it]\n",
      "Loss: 0.02446:  13%|█▎        | 32/250 [01:07<07:26,  2.05s/it]\n",
      "Loss: 0.02643:  13%|█▎        | 32/250 [01:07<07:26,  2.05s/it]\n",
      "Loss: 0.02684:  13%|█▎        | 32/250 [01:08<07:26,  2.05s/it]\n",
      "Loss: 0.02096:  13%|█▎        | 32/250 [01:08<07:26,  2.05s/it]\n",
      "Loss: 0.01883:  13%|█▎        | 32/250 [01:08<07:26,  2.05s/it]\n",
      "Loss: 0.01883:  13%|█▎        | 33/250 [01:08<07:24,  2.05s/it]\n",
      "Loss: 0.02289:  13%|█▎        | 33/250 [01:08<07:24,  2.05s/it]\n",
      "Loss: 0.02202:  13%|█▎        | 33/250 [01:08<07:24,  2.05s/it]\n",
      "Loss: 0.02545:  13%|█▎        | 33/250 [01:08<07:24,  2.05s/it]\n",
      "Loss: 0.02081:  13%|█▎        | 33/250 [01:08<07:24,  2.05s/it]\n",
      "Loss: 0.02205:  13%|█▎        | 33/250 [01:08<07:24,  2.05s/it]\n",
      "Loss: 0.02323:  13%|█▎        | 33/250 [01:08<07:24,  2.05s/it]\n",
      "Loss: 0.02202:  13%|█▎        | 33/250 [01:09<07:24,  2.05s/it]\n",
      "Loss: 0.02519:  13%|█▎        | 33/250 [01:09<07:24,  2.05s/it]\n",
      "Loss: 0.02248:  13%|█▎        | 33/250 [01:09<07:24,  2.05s/it]\n",
      "Loss: 0.02420:  13%|█▎        | 33/250 [01:09<07:24,  2.05s/it]\n",
      "Loss: 0.02276:  13%|█▎        | 33/250 [01:09<07:24,  2.05s/it]\n",
      "Loss: 0.02339:  13%|█▎        | 33/250 [01:09<07:24,  2.05s/it]\n",
      "Loss: 0.02293:  13%|█▎        | 33/250 [01:09<07:24,  2.05s/it]\n",
      "Loss: 0.02472:  13%|█▎        | 33/250 [01:09<07:24,  2.05s/it]\n",
      "Loss: 0.02265:  13%|█▎        | 33/250 [01:09<07:24,  2.05s/it]\n",
      "Loss: 0.02057:  13%|█▎        | 33/250 [01:10<07:24,  2.05s/it]\n",
      "Loss: 0.02407:  13%|█▎        | 33/250 [01:10<07:24,  2.05s/it]\n",
      "Loss: 0.02332:  13%|█▎        | 33/250 [01:10<07:24,  2.05s/it]\n",
      "Loss: 0.01914:  13%|█▎        | 33/250 [01:10<07:24,  2.05s/it]\n",
      "Loss: 0.01914:  14%|█▎        | 34/250 [01:10<07:22,  2.05s/it]\n",
      "Loss: 0.02331:  14%|█▎        | 34/250 [01:10<07:22,  2.05s/it]\n",
      "Loss: 0.02172:  14%|█▎        | 34/250 [01:10<07:22,  2.05s/it]\n",
      "Loss: 0.02544:  14%|█▎        | 34/250 [01:10<07:22,  2.05s/it]\n",
      "Loss: 0.02168:  14%|█▎        | 34/250 [01:10<07:22,  2.05s/it]\n",
      "Loss: 0.02226:  14%|█▎        | 34/250 [01:10<07:22,  2.05s/it]\n",
      "Loss: 0.02071:  14%|█▎        | 34/250 [01:10<07:22,  2.05s/it]\n",
      "Loss: 0.02195:  14%|█▎        | 34/250 [01:11<07:22,  2.05s/it]\n",
      "Loss: 0.02671:  14%|█▎        | 34/250 [01:11<07:22,  2.05s/it]\n",
      "Loss: 0.02210:  14%|█▎        | 34/250 [01:11<07:22,  2.05s/it]\n",
      "Loss: 0.02307:  14%|█▎        | 34/250 [01:11<07:22,  2.05s/it]\n",
      "Loss: 0.02563:  14%|█▎        | 34/250 [01:11<07:22,  2.05s/it]\n",
      "Loss: 0.02256:  14%|█▎        | 34/250 [01:11<07:22,  2.05s/it]\n",
      "Loss: 0.02314:  14%|█▎        | 34/250 [01:11<07:22,  2.05s/it]\n",
      "Loss: 0.02473:  14%|█▎        | 34/250 [01:11<07:22,  2.05s/it]\n",
      "Loss: 0.02283:  14%|█▎        | 34/250 [01:11<07:22,  2.05s/it]\n",
      "Loss: 0.02069:  14%|█▎        | 34/250 [01:12<07:22,  2.05s/it]\n",
      "Loss: 0.02403:  14%|█▎        | 34/250 [01:12<07:22,  2.05s/it]\n",
      "Loss: 0.02348:  14%|█▎        | 34/250 [01:12<07:22,  2.05s/it]\n",
      "Loss: 0.01867:  14%|█▎        | 34/250 [01:12<07:22,  2.05s/it]\n",
      "Loss: 0.01867:  14%|█▍        | 35/250 [01:12<07:20,  2.05s/it]\n",
      "Loss: 0.02422:  14%|█▍        | 35/250 [01:12<07:20,  2.05s/it]\n",
      "Loss: 0.02139:  14%|█▍        | 35/250 [01:12<07:20,  2.05s/it]\n",
      "Loss: 0.02562:  14%|█▍        | 35/250 [01:12<07:20,  2.05s/it]\n",
      "Loss: 0.02162:  14%|█▍        | 35/250 [01:12<07:20,  2.05s/it]\n",
      "Loss: 0.02318:  14%|█▍        | 35/250 [01:12<07:20,  2.05s/it]\n",
      "Loss: 0.02183:  14%|█▍        | 35/250 [01:12<07:20,  2.05s/it]\n",
      "Loss: 0.02162:  14%|█▍        | 35/250 [01:13<07:20,  2.05s/it]\n",
      "Loss: 0.02544:  14%|█▍        | 35/250 [01:13<07:20,  2.05s/it]\n",
      "Loss: 0.02264:  14%|█▍        | 35/250 [01:13<07:20,  2.05s/it]\n",
      "Loss: 0.02379:  14%|█▍        | 35/250 [01:13<07:20,  2.05s/it]\n",
      "Loss: 0.02284:  14%|█▍        | 35/250 [01:13<07:20,  2.05s/it]\n",
      "Loss: 0.02477:  14%|█▍        | 35/250 [01:13<07:20,  2.05s/it]\n",
      "Loss: 0.02228:  14%|█▍        | 35/250 [01:13<07:20,  2.05s/it]\n",
      "Loss: 0.02253:  14%|█▍        | 35/250 [01:13<07:20,  2.05s/it]\n",
      "Loss: 0.02060:  14%|█▍        | 35/250 [01:14<07:20,  2.05s/it]\n",
      "Loss: 0.02089:  14%|█▍        | 35/250 [01:14<07:20,  2.05s/it]\n",
      "Loss: 0.02404:  14%|█▍        | 35/250 [01:14<07:20,  2.05s/it]\n",
      "Loss: 0.01892:  14%|█▍        | 35/250 [01:14<07:20,  2.05s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.37it/s]\u001b[A\n",
      "Loss: 0.01892:  14%|█▍        | 36/250 [01:14<07:17,  2.04s/it]\n",
      "Loss: 0.02281:  14%|█▍        | 36/250 [01:14<07:17,  2.04s/it]\n",
      "Loss: 0.02202:  14%|█▍        | 36/250 [01:14<07:17,  2.04s/it]\n",
      "Loss: 0.02499:  14%|█▍        | 36/250 [01:14<07:17,  2.04s/it]\n",
      "Loss: 0.02328:  14%|█▍        | 36/250 [01:14<07:17,  2.04s/it]\n",
      "Loss: 0.02124:  14%|█▍        | 36/250 [01:14<07:17,  2.04s/it]\n",
      "Loss: 0.02423:  14%|█▍        | 36/250 [01:15<07:17,  2.04s/it]\n",
      "Loss: 0.02268:  14%|█▍        | 36/250 [01:15<07:17,  2.04s/it]\n",
      "Loss: 0.02455:  14%|█▍        | 36/250 [01:15<07:17,  2.04s/it]\n",
      "Loss: 0.02267:  14%|█▍        | 36/250 [01:15<07:17,  2.04s/it]\n",
      "Loss: 0.02211:  14%|█▍        | 36/250 [01:15<07:17,  2.04s/it]\n",
      "Loss: 0.02179:  14%|█▍        | 36/250 [01:15<07:17,  2.04s/it]\n",
      "Loss: 0.02249:  14%|█▍        | 36/250 [01:15<07:17,  2.04s/it]\n",
      "Loss: 0.02205:  14%|█▍        | 36/250 [01:15<07:17,  2.04s/it]\n",
      "Loss: 0.02621:  14%|█▍        | 36/250 [01:15<07:17,  2.04s/it]\n",
      "Loss: 0.02156:  14%|█▍        | 36/250 [01:16<07:17,  2.04s/it]\n",
      "Loss: 0.02188:  14%|█▍        | 36/250 [01:16<07:17,  2.04s/it]\n",
      "Loss: 0.02370:  14%|█▍        | 36/250 [01:16<07:17,  2.04s/it]\n",
      "Loss: 0.02165:  14%|█▍        | 36/250 [01:16<07:17,  2.04s/it]\n",
      "Loss: 0.01892:  14%|█▍        | 36/250 [01:16<07:17,  2.04s/it]\n",
      "Loss: 0.01892:  15%|█▍        | 37/250 [01:16<07:16,  2.05s/it]\n",
      "Loss: 0.02304:  15%|█▍        | 37/250 [01:16<07:16,  2.05s/it]\n",
      "Loss: 0.02010:  15%|█▍        | 37/250 [01:16<07:16,  2.05s/it]\n",
      "Loss: 0.02047:  15%|█▍        | 37/250 [01:16<07:16,  2.05s/it]\n",
      "Loss: 0.02384:  15%|█▍        | 37/250 [01:16<07:16,  2.05s/it]\n",
      "Loss: 0.01891:  15%|█▍        | 37/250 [01:16<07:16,  2.05s/it]\n",
      "Loss: 0.02363:  15%|█▍        | 37/250 [01:17<07:16,  2.05s/it]\n",
      "Loss: 0.02168:  15%|█▍        | 37/250 [01:17<07:16,  2.05s/it]\n",
      "Loss: 0.02350:  15%|█▍        | 37/250 [01:17<07:16,  2.05s/it]\n",
      "Loss: 0.02399:  15%|█▍        | 37/250 [01:17<07:16,  2.05s/it]\n",
      "Loss: 0.02480:  15%|█▍        | 37/250 [01:17<07:16,  2.05s/it]\n",
      "Loss: 0.02307:  15%|█▍        | 37/250 [01:17<07:16,  2.05s/it]\n",
      "Loss: 0.02319:  15%|█▍        | 37/250 [01:17<07:16,  2.05s/it]\n",
      "Loss: 0.02194:  15%|█▍        | 37/250 [01:17<07:16,  2.05s/it]\n",
      "Loss: 0.02112:  15%|█▍        | 37/250 [01:17<07:16,  2.05s/it]\n",
      "Loss: 0.02162:  15%|█▍        | 37/250 [01:18<07:16,  2.05s/it]\n",
      "Loss: 0.02191:  15%|█▍        | 37/250 [01:18<07:16,  2.05s/it]\n",
      "Loss: 0.02353:  15%|█▍        | 37/250 [01:18<07:16,  2.05s/it]\n",
      "Loss: 0.02274:  15%|█▍        | 37/250 [01:18<07:16,  2.05s/it]\n",
      "Loss: 0.01984:  15%|█▍        | 37/250 [01:18<07:16,  2.05s/it]\n",
      "Loss: 0.01984:  15%|█▌        | 38/250 [01:18<07:15,  2.06s/it]\n",
      "Loss: 0.02306:  15%|█▌        | 38/250 [01:18<07:15,  2.06s/it]\n",
      "Loss: 0.02183:  15%|█▌        | 38/250 [01:18<07:15,  2.06s/it]\n",
      "Loss: 0.02473:  15%|█▌        | 38/250 [01:18<07:15,  2.06s/it]\n",
      "Loss: 0.02175:  15%|█▌        | 38/250 [01:18<07:15,  2.06s/it]\n",
      "Loss: 0.02135:  15%|█▌        | 38/250 [01:19<07:15,  2.06s/it]\n",
      "Loss: 0.02188:  15%|█▌        | 38/250 [01:19<07:15,  2.06s/it]\n",
      "Loss: 0.02484:  15%|█▌        | 38/250 [01:19<07:15,  2.06s/it]\n",
      "Loss: 0.02350:  15%|█▌        | 38/250 [01:19<07:15,  2.06s/it]\n",
      "Loss: 0.02011:  15%|█▌        | 38/250 [01:19<07:15,  2.06s/it]\n",
      "Loss: 0.02342:  15%|█▌        | 38/250 [01:19<07:15,  2.06s/it]\n",
      "Loss: 0.02263:  15%|█▌        | 38/250 [01:19<07:15,  2.06s/it]\n",
      "Loss: 0.02490:  15%|█▌        | 38/250 [01:19<07:15,  2.06s/it]\n",
      "Loss: 0.02197:  15%|█▌        | 38/250 [01:19<07:15,  2.06s/it]\n",
      "Loss: 0.02195:  15%|█▌        | 38/250 [01:20<07:15,  2.06s/it]\n",
      "Loss: 0.02055:  15%|█▌        | 38/250 [01:20<07:15,  2.06s/it]\n",
      "Loss: 0.02030:  15%|█▌        | 38/250 [01:20<07:15,  2.06s/it]\n",
      "Loss: 0.02266:  15%|█▌        | 38/250 [01:20<07:15,  2.06s/it]\n",
      "Loss: 0.01881:  15%|█▌        | 38/250 [01:20<07:15,  2.06s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 60.12it/s]\u001b[A\n",
      "Loss: 0.01881:  16%|█▌        | 39/250 [01:20<07:13,  2.05s/it]\n",
      "Loss: 0.02267:  16%|█▌        | 39/250 [01:20<07:13,  2.05s/it]\n",
      "Loss: 0.02184:  16%|█▌        | 39/250 [01:20<07:13,  2.05s/it]\n",
      "Loss: 0.02459:  16%|█▌        | 39/250 [01:20<07:13,  2.05s/it]\n",
      "Loss: 0.02026:  16%|█▌        | 39/250 [01:20<07:13,  2.05s/it]\n",
      "Loss: 0.02168:  16%|█▌        | 39/250 [01:21<07:13,  2.05s/it]\n",
      "Loss: 0.02301:  16%|█▌        | 39/250 [01:21<07:13,  2.05s/it]\n",
      "Loss: 0.02130:  16%|█▌        | 39/250 [01:21<07:13,  2.05s/it]\n",
      "Loss: 0.02239:  16%|█▌        | 39/250 [01:21<07:13,  2.05s/it]\n",
      "Loss: 0.02412:  16%|█▌        | 39/250 [01:21<07:13,  2.05s/it]\n",
      "Loss: 0.02257:  16%|█▌        | 39/250 [01:21<07:13,  2.05s/it]\n",
      "Loss: 0.02199:  16%|█▌        | 39/250 [01:21<07:13,  2.05s/it]\n",
      "Loss: 0.02619:  16%|█▌        | 39/250 [01:21<07:13,  2.05s/it]\n",
      "Loss: 0.02178:  16%|█▌        | 39/250 [01:21<07:13,  2.05s/it]\n",
      "Loss: 0.02499:  16%|█▌        | 39/250 [01:22<07:13,  2.05s/it]\n",
      "Loss: 0.02077:  16%|█▌        | 39/250 [01:22<07:13,  2.05s/it]\n",
      "Loss: 0.02097:  16%|█▌        | 39/250 [01:22<07:13,  2.05s/it]\n",
      "Loss: 0.02147:  16%|█▌        | 39/250 [01:22<07:13,  2.05s/it]\n",
      "Loss: 0.02228:  16%|█▌        | 39/250 [01:22<07:13,  2.05s/it]\n",
      "Loss: 0.01873:  16%|█▌        | 39/250 [01:22<07:13,  2.05s/it]\n",
      "Loss: 0.01873:  16%|█▌        | 40/250 [01:22<07:10,  2.05s/it]\n",
      "Loss: 0.02295:  16%|█▌        | 40/250 [01:22<07:10,  2.05s/it]\n",
      "Loss: 0.01965:  16%|█▌        | 40/250 [01:22<07:10,  2.05s/it]\n",
      "Loss: 0.02428:  16%|█▌        | 40/250 [01:22<07:10,  2.05s/it]\n",
      "Loss: 0.02091:  16%|█▌        | 40/250 [01:23<07:10,  2.05s/it]\n",
      "Loss: 0.01847:  16%|█▌        | 40/250 [01:23<07:10,  2.05s/it]\n",
      "Loss: 0.02281:  16%|█▌        | 40/250 [01:23<07:10,  2.05s/it]\n",
      "Loss: 0.02171:  16%|█▌        | 40/250 [01:23<07:10,  2.05s/it]\n",
      "Loss: 0.02138:  16%|█▌        | 40/250 [01:23<07:10,  2.05s/it]\n",
      "Loss: 0.02288:  16%|█▌        | 40/250 [01:23<07:10,  2.05s/it]\n",
      "Loss: 0.02230:  16%|█▌        | 40/250 [01:23<07:10,  2.05s/it]\n",
      "Loss: 0.02469:  16%|█▌        | 40/250 [01:23<07:10,  2.05s/it]\n",
      "Loss: 0.02306:  16%|█▌        | 40/250 [01:23<07:10,  2.05s/it]\n",
      "Loss: 0.02276:  16%|█▌        | 40/250 [01:24<07:10,  2.05s/it]\n",
      "Loss: 0.02430:  16%|█▌        | 40/250 [01:24<07:10,  2.05s/it]\n",
      "Loss: 0.02235:  16%|█▌        | 40/250 [01:24<07:10,  2.05s/it]\n",
      "Loss: 0.02132:  16%|█▌        | 40/250 [01:24<07:10,  2.05s/it]\n",
      "Loss: 0.02357:  16%|█▌        | 40/250 [01:24<07:10,  2.05s/it]\n",
      "Loss: 0.02318:  16%|█▌        | 40/250 [01:24<07:10,  2.05s/it]\n",
      "Loss: 0.01860:  16%|█▌        | 40/250 [01:24<07:10,  2.05s/it]\n",
      "Loss: 0.01860:  16%|█▋        | 41/250 [01:24<07:08,  2.05s/it]\n",
      "Loss: 0.02187:  16%|█▋        | 41/250 [01:24<07:08,  2.05s/it]\n",
      "Loss: 0.01979:  16%|█▋        | 41/250 [01:24<07:08,  2.05s/it]\n",
      "Loss: 0.02062:  16%|█▋        | 41/250 [01:24<07:08,  2.05s/it]\n",
      "Loss: 0.02387:  16%|█▋        | 41/250 [01:25<07:08,  2.05s/it]\n",
      "Loss: 0.01879:  16%|█▋        | 41/250 [01:25<07:08,  2.05s/it]\n",
      "Loss: 0.02278:  16%|█▋        | 41/250 [01:25<07:08,  2.05s/it]\n",
      "Loss: 0.02053:  16%|█▋        | 41/250 [01:25<07:08,  2.05s/it]\n",
      "Loss: 0.02133:  16%|█▋        | 41/250 [01:25<07:08,  2.05s/it]\n",
      "Loss: 0.02365:  16%|█▋        | 41/250 [01:25<07:08,  2.05s/it]\n",
      "Loss: 0.02261:  16%|█▋        | 41/250 [01:25<07:08,  2.05s/it]\n",
      "Loss: 0.02058:  16%|█▋        | 41/250 [01:25<07:08,  2.05s/it]\n",
      "Loss: 0.02168:  16%|█▋        | 41/250 [01:25<07:08,  2.05s/it]\n",
      "Loss: 0.02492:  16%|█▋        | 41/250 [01:26<07:08,  2.05s/it]\n",
      "Loss: 0.01944:  16%|█▋        | 41/250 [01:26<07:08,  2.05s/it]\n",
      "Loss: 0.02444:  16%|█▋        | 41/250 [01:26<07:08,  2.05s/it]\n",
      "Loss: 0.02596:  16%|█▋        | 41/250 [01:26<07:08,  2.05s/it]\n",
      "Loss: 0.02492:  16%|█▋        | 41/250 [01:26<07:08,  2.05s/it]\n",
      "Loss: 0.02142:  16%|█▋        | 41/250 [01:26<07:08,  2.05s/it]\n",
      "Loss: 0.01863:  16%|█▋        | 41/250 [01:26<07:08,  2.05s/it]\n",
      "Loss: 0.01863:  17%|█▋        | 42/250 [01:26<07:06,  2.05s/it]\n",
      "Loss: 0.02255:  17%|█▋        | 42/250 [01:26<07:06,  2.05s/it]\n",
      "Loss: 0.02141:  17%|█▋        | 42/250 [01:26<07:06,  2.05s/it]\n",
      "Loss: 0.02452:  17%|█▋        | 42/250 [01:27<07:06,  2.05s/it]\n",
      "Loss: 0.02180:  17%|█▋        | 42/250 [01:27<07:06,  2.05s/it]\n",
      "Loss: 0.02154:  17%|█▋        | 42/250 [01:27<07:06,  2.05s/it]\n",
      "Loss: 0.02113:  17%|█▋        | 42/250 [01:27<07:06,  2.05s/it]\n",
      "Loss: 0.02250:  17%|█▋        | 42/250 [01:27<07:06,  2.05s/it]\n",
      "Loss: 0.02299:  17%|█▋        | 42/250 [01:27<07:06,  2.05s/it]\n",
      "Loss: 0.02213:  17%|█▋        | 42/250 [01:27<07:06,  2.05s/it]\n",
      "Loss: 0.02180:  17%|█▋        | 42/250 [01:27<07:06,  2.05s/it]\n",
      "Loss: 0.02083:  17%|█▋        | 42/250 [01:27<07:06,  2.05s/it]\n",
      "Loss: 0.02216:  17%|█▋        | 42/250 [01:28<07:06,  2.05s/it]\n",
      "Loss: 0.02244:  17%|█▋        | 42/250 [01:28<07:06,  2.05s/it]\n",
      "Loss: 0.02124:  17%|█▋        | 42/250 [01:28<07:06,  2.05s/it]\n",
      "Loss: 0.02431:  17%|█▋        | 42/250 [01:28<07:06,  2.05s/it]\n",
      "Loss: 0.02635:  17%|█▋        | 42/250 [01:28<07:06,  2.05s/it]\n",
      "Loss: 0.02516:  17%|█▋        | 42/250 [01:28<07:06,  2.05s/it]\n",
      "Loss: 0.02069:  17%|█▋        | 42/250 [01:28<07:06,  2.05s/it]\n",
      "Loss: 0.01986:  17%|█▋        | 42/250 [01:28<07:06,  2.05s/it]\n",
      "Loss: 0.01986:  17%|█▋        | 43/250 [01:28<07:03,  2.05s/it]\n",
      "Loss: 0.02260:  17%|█▋        | 43/250 [01:28<07:03,  2.05s/it]\n",
      "Loss: 0.02131:  17%|█▋        | 43/250 [01:28<07:03,  2.05s/it]\n",
      "Loss: 0.02422:  17%|█▋        | 43/250 [01:29<07:03,  2.05s/it]\n",
      "Loss: 0.02426:  17%|█▋        | 43/250 [01:29<07:03,  2.05s/it]\n",
      "Loss: 0.01898:  17%|█▋        | 43/250 [01:29<07:03,  2.05s/it]\n",
      "Loss: 0.02320:  17%|█▋        | 43/250 [01:29<07:03,  2.05s/it]\n",
      "Loss: 0.02367:  17%|█▋        | 43/250 [01:29<07:03,  2.05s/it]\n",
      "Loss: 0.02294:  17%|█▋        | 43/250 [01:29<07:03,  2.05s/it]\n",
      "Loss: 0.02324:  17%|█▋        | 43/250 [01:29<07:03,  2.05s/it]\n",
      "Loss: 0.02241:  17%|█▋        | 43/250 [01:29<07:03,  2.05s/it]\n",
      "Loss: 0.02060:  17%|█▋        | 43/250 [01:29<07:03,  2.05s/it]\n",
      "Loss: 0.02201:  17%|█▋        | 43/250 [01:30<07:03,  2.05s/it]\n",
      "Loss: 0.02324:  17%|█▋        | 43/250 [01:30<07:03,  2.05s/it]\n",
      "Loss: 0.01977:  17%|█▋        | 43/250 [01:30<07:03,  2.05s/it]\n",
      "Loss: 0.02366:  17%|█▋        | 43/250 [01:30<07:03,  2.05s/it]\n",
      "Loss: 0.02559:  17%|█▋        | 43/250 [01:30<07:03,  2.05s/it]\n",
      "Loss: 0.02499:  17%|█▋        | 43/250 [01:30<07:03,  2.05s/it]\n",
      "Loss: 0.02035:  17%|█▋        | 43/250 [01:30<07:03,  2.05s/it]\n",
      "Loss: 0.01880:  17%|█▋        | 43/250 [01:30<07:03,  2.05s/it]\n",
      "Loss: 0.01880:  18%|█▊        | 44/250 [01:30<07:02,  2.05s/it]\n",
      "Loss: 0.02275:  18%|█▊        | 44/250 [01:30<07:02,  2.05s/it]\n",
      "Loss: 0.01976:  18%|█▊        | 44/250 [01:31<07:02,  2.05s/it]\n",
      "Loss: 0.01969:  18%|█▊        | 44/250 [01:31<07:02,  2.05s/it]\n",
      "Loss: 0.02473:  18%|█▊        | 44/250 [01:31<07:02,  2.05s/it]\n",
      "Loss: 0.02347:  18%|█▊        | 44/250 [01:31<07:02,  2.05s/it]\n",
      "Loss: 0.02264:  18%|█▊        | 44/250 [01:31<07:02,  2.05s/it]\n",
      "Loss: 0.02442:  18%|█▊        | 44/250 [01:31<07:02,  2.05s/it]\n",
      "Loss: 0.02251:  18%|█▊        | 44/250 [01:31<07:02,  2.05s/it]\n",
      "Loss: 0.02309:  18%|█▊        | 44/250 [01:31<07:02,  2.05s/it]\n",
      "Loss: 0.02142:  18%|█▊        | 44/250 [01:31<07:02,  2.05s/it]\n",
      "Loss: 0.02160:  18%|█▊        | 44/250 [01:31<07:02,  2.05s/it]\n",
      "Loss: 0.02560:  18%|█▊        | 44/250 [01:32<07:02,  2.05s/it]\n",
      "Loss: 0.02151:  18%|█▊        | 44/250 [01:32<07:02,  2.05s/it]\n",
      "Loss: 0.02437:  18%|█▊        | 44/250 [01:32<07:02,  2.05s/it]\n",
      "Loss: 0.02144:  18%|█▊        | 44/250 [01:32<07:02,  2.05s/it]\n",
      "Loss: 0.02080:  18%|█▊        | 44/250 [01:32<07:02,  2.05s/it]\n",
      "Loss: 0.02066:  18%|█▊        | 44/250 [01:32<07:02,  2.05s/it]\n",
      "Loss: 0.02220:  18%|█▊        | 44/250 [01:32<07:02,  2.05s/it]\n",
      "Loss: 0.01853:  18%|█▊        | 44/250 [01:32<07:02,  2.05s/it]\n",
      "Loss: 0.01853:  18%|█▊        | 45/250 [01:32<07:01,  2.05s/it]\n",
      "Loss: 0.02246:  18%|█▊        | 45/250 [01:32<07:01,  2.05s/it]\n",
      "Loss: 0.02135:  18%|█▊        | 45/250 [01:33<07:01,  2.05s/it]\n",
      "Loss: 0.02540:  18%|█▊        | 45/250 [01:33<07:01,  2.05s/it]\n",
      "Loss: 0.02153:  18%|█▊        | 45/250 [01:33<07:01,  2.05s/it]\n",
      "Loss: 0.02135:  18%|█▊        | 45/250 [01:33<07:01,  2.05s/it]\n",
      "Loss: 0.02138:  18%|█▊        | 45/250 [01:33<07:01,  2.05s/it]\n",
      "Loss: 0.02229:  18%|█▊        | 45/250 [01:33<07:01,  2.05s/it]\n",
      "Loss: 0.02369:  18%|█▊        | 45/250 [01:33<07:01,  2.05s/it]\n",
      "Loss: 0.02013:  18%|█▊        | 45/250 [01:33<07:01,  2.05s/it]\n",
      "Loss: 0.02384:  18%|█▊        | 45/250 [01:33<07:01,  2.05s/it]\n",
      "Loss: 0.02473:  18%|█▊        | 45/250 [01:34<07:01,  2.05s/it]\n",
      "Loss: 0.02266:  18%|█▊        | 45/250 [01:34<07:01,  2.05s/it]\n",
      "Loss: 0.02222:  18%|█▊        | 45/250 [01:34<07:01,  2.05s/it]\n",
      "Loss: 0.02100:  18%|█▊        | 45/250 [01:34<07:01,  2.05s/it]\n",
      "Loss: 0.02168:  18%|█▊        | 45/250 [01:34<07:01,  2.05s/it]\n",
      "Loss: 0.02359:  18%|█▊        | 45/250 [01:34<07:01,  2.05s/it]\n",
      "Loss: 0.02177:  18%|█▊        | 45/250 [01:34<07:01,  2.05s/it]\n",
      "Loss: 0.01852:  18%|█▊        | 45/250 [01:34<07:01,  2.05s/it]\n",
      "Loss: 0.01852:  18%|█▊        | 46/250 [01:34<06:56,  2.04s/it]\n",
      "Loss: 0.02233:  18%|█▊        | 46/250 [01:34<06:56,  2.04s/it]\n",
      "Loss: 0.02115:  18%|█▊        | 46/250 [01:35<06:56,  2.04s/it]\n",
      "Loss: 0.02462:  18%|█▊        | 46/250 [01:35<06:56,  2.04s/it]\n",
      "Loss: 0.02168:  18%|█▊        | 46/250 [01:35<06:56,  2.04s/it]\n",
      "Loss: 0.02100:  18%|█▊        | 46/250 [01:35<06:56,  2.04s/it]\n",
      "Loss: 0.02219:  18%|█▊        | 46/250 [01:35<06:56,  2.04s/it]\n",
      "Loss: 0.02215:  18%|█▊        | 46/250 [01:35<06:56,  2.04s/it]\n",
      "Loss: 0.02254:  18%|█▊        | 46/250 [01:35<06:56,  2.04s/it]\n",
      "Loss: 0.02000:  18%|█▊        | 46/250 [01:35<06:56,  2.04s/it]\n",
      "Loss: 0.02356:  18%|█▊        | 46/250 [01:35<06:56,  2.04s/it]\n",
      "Loss: 0.02475:  18%|█▊        | 46/250 [01:36<06:56,  2.04s/it]\n",
      "Loss: 0.02257:  18%|█▊        | 46/250 [01:36<06:56,  2.04s/it]\n",
      "Loss: 0.02253:  18%|█▊        | 46/250 [01:36<06:56,  2.04s/it]\n",
      "Loss: 0.02166:  18%|█▊        | 46/250 [01:36<06:56,  2.04s/it]\n",
      "Loss: 0.02185:  18%|█▊        | 46/250 [01:36<06:56,  2.04s/it]\n",
      "Loss: 0.02409:  18%|█▊        | 46/250 [01:36<06:56,  2.04s/it]\n",
      "Loss: 0.02189:  18%|█▊        | 46/250 [01:36<06:56,  2.04s/it]\n",
      "Loss: 0.01871:  18%|█▊        | 46/250 [01:36<06:56,  2.04s/it]\n",
      "Loss: 0.01871:  19%|█▉        | 47/250 [01:36<06:52,  2.03s/it]\n",
      "Loss: 0.02110:  19%|█▉        | 47/250 [01:36<06:52,  2.03s/it]\n",
      "Loss: 0.01990:  19%|█▉        | 47/250 [01:37<06:52,  2.03s/it]\n",
      "Loss: 0.02422:  19%|█▉        | 47/250 [01:37<06:52,  2.03s/it]\n",
      "Loss: 0.02016:  19%|█▉        | 47/250 [01:37<06:52,  2.03s/it]\n",
      "Loss: 0.02243:  19%|█▉        | 47/250 [01:37<06:52,  2.03s/it]\n",
      "Loss: 0.02296:  19%|█▉        | 47/250 [01:37<06:52,  2.03s/it]\n",
      "Loss: 0.02057:  19%|█▉        | 47/250 [01:37<06:52,  2.03s/it]\n",
      "Loss: 0.02112:  19%|█▉        | 47/250 [01:37<06:52,  2.03s/it]\n",
      "Loss: 0.02285:  19%|█▉        | 47/250 [01:37<06:52,  2.03s/it]\n",
      "Loss: 0.02190:  19%|█▉        | 47/250 [01:37<06:52,  2.03s/it]\n",
      "Loss: 0.02542:  19%|█▉        | 47/250 [01:38<06:52,  2.03s/it]\n",
      "Loss: 0.02224:  19%|█▉        | 47/250 [01:38<06:52,  2.03s/it]\n",
      "Loss: 0.02334:  19%|█▉        | 47/250 [01:38<06:52,  2.03s/it]\n",
      "Loss: 0.02424:  19%|█▉        | 47/250 [01:38<06:52,  2.03s/it]\n",
      "Loss: 0.02186:  19%|█▉        | 47/250 [01:38<06:52,  2.03s/it]\n",
      "Loss: 0.01986:  19%|█▉        | 47/250 [01:38<06:52,  2.03s/it]\n",
      "Loss: 0.02350:  19%|█▉        | 47/250 [01:38<06:52,  2.03s/it]\n",
      "Loss: 0.02311:  19%|█▉        | 47/250 [01:38<06:52,  2.03s/it]\n",
      "Loss: 0.01879:  19%|█▉        | 47/250 [01:38<06:52,  2.03s/it]\n",
      "Loss: 0.01879:  19%|█▉        | 48/250 [01:38<06:52,  2.04s/it]\n",
      "Loss: 0.02251:  19%|█▉        | 48/250 [01:39<06:52,  2.04s/it]\n",
      "Loss: 0.01928:  19%|█▉        | 48/250 [01:39<06:52,  2.04s/it]\n",
      "Loss: 0.02446:  19%|█▉        | 48/250 [01:39<06:52,  2.04s/it]\n",
      "Loss: 0.02082:  19%|█▉        | 48/250 [01:39<06:52,  2.04s/it]\n",
      "Loss: 0.01819:  19%|█▉        | 48/250 [01:39<06:52,  2.04s/it]\n",
      "Loss: 0.02267:  19%|█▉        | 48/250 [01:39<06:52,  2.04s/it]\n",
      "Loss: 0.02256:  19%|█▉        | 48/250 [01:39<06:52,  2.04s/it]\n",
      "Loss: 0.02523:  19%|█▉        | 48/250 [01:39<06:52,  2.04s/it]\n",
      "Loss: 0.02191:  19%|█▉        | 48/250 [01:39<06:52,  2.04s/it]\n",
      "Loss: 0.02399:  19%|█▉        | 48/250 [01:40<06:52,  2.04s/it]\n",
      "Loss: 0.02240:  19%|█▉        | 48/250 [01:40<06:52,  2.04s/it]\n",
      "Loss: 0.01983:  19%|█▉        | 48/250 [01:40<06:52,  2.04s/it]\n",
      "Loss: 0.02514:  19%|█▉        | 48/250 [01:40<06:52,  2.04s/it]\n",
      "Loss: 0.02213:  19%|█▉        | 48/250 [01:40<06:52,  2.04s/it]\n",
      "Loss: 0.02080:  19%|█▉        | 48/250 [01:40<06:52,  2.04s/it]\n",
      "Loss: 0.02513:  19%|█▉        | 48/250 [01:40<06:52,  2.04s/it]\n",
      "Loss: 0.02005:  19%|█▉        | 48/250 [01:40<06:52,  2.04s/it]\n",
      "Loss: 0.02170:  19%|█▉        | 48/250 [01:40<06:52,  2.04s/it]\n",
      "Loss: 0.01837:  19%|█▉        | 48/250 [01:40<06:52,  2.04s/it]\n",
      "Loss: 0.01837:  20%|█▉        | 49/250 [01:40<06:50,  2.04s/it]\n",
      "Loss: 0.02098:  20%|█▉        | 49/250 [01:41<06:50,  2.04s/it]\n",
      "Loss: 0.01998:  20%|█▉        | 49/250 [01:41<06:50,  2.04s/it]\n",
      "Loss: 0.02396:  20%|█▉        | 49/250 [01:41<06:50,  2.04s/it]\n",
      "Loss: 0.02016:  20%|█▉        | 49/250 [01:41<06:50,  2.04s/it]\n",
      "Loss: 0.02184:  20%|█▉        | 49/250 [01:41<06:50,  2.04s/it]\n",
      "Loss: 0.01996:  20%|█▉        | 49/250 [01:41<06:50,  2.04s/it]\n",
      "Loss: 0.02131:  20%|█▉        | 49/250 [01:41<06:50,  2.04s/it]\n",
      "Loss: 0.02493:  20%|█▉        | 49/250 [01:41<06:50,  2.04s/it]\n",
      "Loss: 0.02171:  20%|█▉        | 49/250 [01:41<06:50,  2.04s/it]\n",
      "Loss: 0.02345:  20%|█▉        | 49/250 [01:42<06:50,  2.04s/it]\n",
      "Loss: 0.02240:  20%|█▉        | 49/250 [01:42<06:50,  2.04s/it]\n",
      "Loss: 0.02471:  20%|█▉        | 49/250 [01:42<06:50,  2.04s/it]\n",
      "Loss: 0.02106:  20%|█▉        | 49/250 [01:42<06:50,  2.04s/it]\n",
      "Loss: 0.02431:  20%|█▉        | 49/250 [01:42<06:50,  2.04s/it]\n",
      "Loss: 0.02210:  20%|█▉        | 49/250 [01:42<06:50,  2.04s/it]\n",
      "Loss: 0.02026:  20%|█▉        | 49/250 [01:42<06:50,  2.04s/it]\n",
      "Loss: 0.02421:  20%|█▉        | 49/250 [01:42<06:50,  2.04s/it]\n",
      "Loss: 0.02293:  20%|█▉        | 49/250 [01:42<06:50,  2.04s/it]\n",
      "Loss: 0.01896:  20%|█▉        | 49/250 [01:43<06:50,  2.04s/it]\n",
      "Loss: 0.01896:  20%|██        | 50/250 [01:43<06:48,  2.04s/it]\n",
      "Loss: 0.02219:  20%|██        | 50/250 [01:43<06:48,  2.04s/it]\n",
      "Loss: 0.01986:  20%|██        | 50/250 [01:43<06:48,  2.04s/it]\n",
      "Loss: 0.02406:  20%|██        | 50/250 [01:43<06:48,  2.04s/it]\n",
      "Loss: 0.02355:  20%|██        | 50/250 [01:43<06:48,  2.04s/it]\n",
      "Loss: 0.01823:  20%|██        | 50/250 [01:43<06:48,  2.04s/it]\n",
      "Loss: 0.02273:  20%|██        | 50/250 [01:43<06:48,  2.04s/it]\n",
      "Loss: 0.02399:  20%|██        | 50/250 [01:43<06:48,  2.04s/it]\n",
      "Loss: 0.02237:  20%|██        | 50/250 [01:43<06:48,  2.04s/it]\n",
      "Loss: 0.02366:  20%|██        | 50/250 [01:43<06:48,  2.04s/it]\n",
      "Loss: 0.02141:  20%|██        | 50/250 [01:44<06:48,  2.04s/it]\n",
      "Loss: 0.02116:  20%|██        | 50/250 [01:44<06:48,  2.04s/it]\n",
      "Loss: 0.02483:  20%|██        | 50/250 [01:44<06:48,  2.04s/it]\n",
      "Loss: 0.02155:  20%|██        | 50/250 [01:44<06:48,  2.04s/it]\n",
      "Loss: 0.02444:  20%|██        | 50/250 [01:44<06:48,  2.04s/it]\n",
      "Loss: 0.02027:  20%|██        | 50/250 [01:44<06:48,  2.04s/it]\n",
      "Loss: 0.02114:  20%|██        | 50/250 [01:44<06:48,  2.04s/it]\n",
      "Loss: 0.02074:  20%|██        | 50/250 [01:44<06:48,  2.04s/it]\n",
      "Loss: 0.02158:  20%|██        | 50/250 [01:45<06:48,  2.04s/it]\n",
      "Loss: 0.01847:  20%|██        | 50/250 [01:45<06:48,  2.04s/it]\n",
      "Loss: 0.01847:  20%|██        | 51/250 [01:45<06:48,  2.05s/it]\n",
      "Loss: 0.02191:  20%|██        | 51/250 [01:45<06:48,  2.05s/it]\n",
      "Loss: 0.01987:  20%|██        | 51/250 [01:45<06:48,  2.05s/it]\n",
      "Loss: 0.01970:  20%|██        | 51/250 [01:45<06:48,  2.05s/it]\n",
      "Loss: 0.02100:  20%|██        | 51/250 [01:45<06:48,  2.05s/it]\n",
      "Loss: 0.02329:  20%|██        | 51/250 [01:45<06:48,  2.05s/it]\n",
      "Loss: 0.02113:  20%|██        | 51/250 [01:45<06:48,  2.05s/it]\n",
      "Loss: 0.02172:  20%|██        | 51/250 [01:45<06:48,  2.05s/it]\n",
      "Loss: 0.02503:  20%|██        | 51/250 [01:45<06:48,  2.05s/it]\n",
      "Loss: 0.02185:  20%|██        | 51/250 [01:46<06:48,  2.05s/it]\n",
      "Loss: 0.02411:  20%|██        | 51/250 [01:46<06:48,  2.05s/it]\n",
      "Loss: 0.02264:  20%|██        | 51/250 [01:46<06:48,  2.05s/it]\n",
      "Loss: 0.02456:  20%|██        | 51/250 [01:46<06:48,  2.05s/it]\n",
      "Loss: 0.02264:  20%|██        | 51/250 [01:46<06:48,  2.05s/it]\n",
      "Loss: 0.02067:  20%|██        | 51/250 [01:46<06:48,  2.05s/it]\n",
      "Loss: 0.02406:  20%|██        | 51/250 [01:46<06:48,  2.05s/it]\n",
      "Loss: 0.02495:  20%|██        | 51/250 [01:46<06:48,  2.05s/it]\n",
      "Loss: 0.02084:  20%|██        | 51/250 [01:46<06:48,  2.05s/it]\n",
      "Loss: 0.01975:  20%|██        | 51/250 [01:47<06:48,  2.05s/it]\n",
      "Loss: 0.02150:  20%|██        | 51/250 [01:47<06:48,  2.05s/it]\n",
      "Loss: 0.01851:  20%|██        | 51/250 [01:47<06:48,  2.05s/it]\n",
      "Loss: 0.01851:  21%|██        | 52/250 [01:47<06:49,  2.07s/it]\n",
      "Loss: 0.02252:  21%|██        | 52/250 [01:47<06:49,  2.07s/it]\n",
      "Loss: 0.02122:  21%|██        | 52/250 [01:47<06:49,  2.07s/it]\n",
      "Loss: 0.02436:  21%|██        | 52/250 [01:47<06:49,  2.07s/it]\n",
      "Loss: 0.02202:  21%|██        | 52/250 [01:47<06:49,  2.07s/it]\n",
      "Loss: 0.02136:  21%|██        | 52/250 [01:47<06:49,  2.07s/it]\n",
      "Loss: 0.01981:  21%|██        | 52/250 [01:47<06:49,  2.07s/it]\n",
      "Loss: 0.02139:  21%|██        | 52/250 [01:47<06:49,  2.07s/it]\n",
      "Loss: 0.02430:  21%|██        | 52/250 [01:48<06:49,  2.07s/it]\n",
      "Loss: 0.02161:  21%|██        | 52/250 [01:48<06:49,  2.07s/it]\n",
      "Loss: 0.02195:  21%|██        | 52/250 [01:48<06:49,  2.07s/it]\n",
      "Loss: 0.02554:  21%|██        | 52/250 [01:48<06:49,  2.07s/it]\n",
      "Loss: 0.02114:  21%|██        | 52/250 [01:48<06:49,  2.07s/it]\n",
      "Loss: 0.02208:  21%|██        | 52/250 [01:48<06:49,  2.07s/it]\n",
      "Loss: 0.02446:  21%|██        | 52/250 [01:48<06:49,  2.07s/it]\n",
      "Loss: 0.02078:  21%|██        | 52/250 [01:48<06:49,  2.07s/it]\n",
      "Loss: 0.02177:  21%|██        | 52/250 [01:48<06:49,  2.07s/it]\n",
      "Loss: 0.02341:  21%|██        | 52/250 [01:49<06:49,  2.07s/it]\n",
      "Loss: 0.02278:  21%|██        | 52/250 [01:49<06:49,  2.07s/it]\n",
      "Loss: 0.01859:  21%|██        | 52/250 [01:49<06:49,  2.07s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 59.67it/s]\u001b[A\n",
      "Loss: 0.01859:  21%|██        | 53/250 [01:49<06:48,  2.07s/it]\n",
      "Loss: 0.02219:  21%|██        | 53/250 [01:49<06:48,  2.07s/it]\n",
      "Loss: 0.01942:  21%|██        | 53/250 [01:49<06:48,  2.07s/it]\n",
      "Loss: 0.01944:  21%|██        | 53/250 [01:49<06:48,  2.07s/it]\n",
      "Loss: 0.02135:  21%|██        | 53/250 [01:49<06:48,  2.07s/it]\n",
      "Loss: 0.02351:  21%|██        | 53/250 [01:49<06:48,  2.07s/it]\n",
      "Loss: 0.02316:  21%|██        | 53/250 [01:49<06:48,  2.07s/it]\n",
      "Loss: 0.02300:  21%|██        | 53/250 [01:50<06:48,  2.07s/it]\n",
      "Loss: 0.02460:  21%|██        | 53/250 [01:50<06:48,  2.07s/it]\n",
      "Loss: 0.02183:  21%|██        | 53/250 [01:50<06:48,  2.07s/it]\n",
      "Loss: 0.02267:  21%|██        | 53/250 [01:50<06:48,  2.07s/it]\n",
      "Loss: 0.02379:  21%|██        | 53/250 [01:50<06:48,  2.07s/it]\n",
      "Loss: 0.02460:  21%|██        | 53/250 [01:50<06:48,  2.07s/it]\n",
      "Loss: 0.02412:  21%|██        | 53/250 [01:50<06:48,  2.07s/it]\n",
      "Loss: 0.02274:  21%|██        | 53/250 [01:50<06:48,  2.07s/it]\n",
      "Loss: 0.01907:  21%|██        | 53/250 [01:50<06:48,  2.07s/it]\n",
      "Loss: 0.02349:  21%|██        | 53/250 [01:50<06:48,  2.07s/it]\n",
      "Loss: 0.02065:  21%|██        | 53/250 [01:51<06:48,  2.07s/it]\n",
      "Loss: 0.02060:  21%|██        | 53/250 [01:51<06:48,  2.07s/it]\n",
      "Loss: 0.02174:  21%|██        | 53/250 [01:51<06:48,  2.07s/it]\n",
      "Loss: 0.01839:  21%|██        | 53/250 [01:51<06:48,  2.07s/it]\n",
      "Loss: 0.01839:  22%|██▏       | 54/250 [01:51<06:48,  2.08s/it]\n",
      "Loss: 0.02284:  22%|██▏       | 54/250 [01:51<06:48,  2.08s/it]\n",
      "Loss: 0.01917:  22%|██▏       | 54/250 [01:51<06:48,  2.08s/it]\n",
      "Loss: 0.01928:  22%|██▏       | 54/250 [01:51<06:48,  2.08s/it]\n",
      "Loss: 0.02199:  22%|██▏       | 54/250 [01:51<06:48,  2.08s/it]\n",
      "Loss: 0.02283:  22%|██▏       | 54/250 [01:51<06:48,  2.08s/it]\n",
      "Loss: 0.02146:  22%|██▏       | 54/250 [01:52<06:48,  2.08s/it]\n",
      "Loss: 0.02193:  22%|██▏       | 54/250 [01:52<06:48,  2.08s/it]\n",
      "Loss: 0.02406:  22%|██▏       | 54/250 [01:52<06:48,  2.08s/it]\n",
      "Loss: 0.02209:  22%|██▏       | 54/250 [01:52<06:48,  2.08s/it]\n",
      "Loss: 0.02280:  22%|██▏       | 54/250 [01:52<06:48,  2.08s/it]\n",
      "Loss: 0.02332:  22%|██▏       | 54/250 [01:52<06:48,  2.08s/it]\n",
      "Loss: 0.02513:  22%|██▏       | 54/250 [01:52<06:48,  2.08s/it]\n",
      "Loss: 0.02407:  22%|██▏       | 54/250 [01:52<06:48,  2.08s/it]\n",
      "Loss: 0.02255:  22%|██▏       | 54/250 [01:52<06:48,  2.08s/it]\n",
      "Loss: 0.01924:  22%|██▏       | 54/250 [01:52<06:48,  2.08s/it]\n",
      "Loss: 0.02417:  22%|██▏       | 54/250 [01:53<06:48,  2.08s/it]\n",
      "Loss: 0.02528:  22%|██▏       | 54/250 [01:53<06:48,  2.08s/it]\n",
      "Loss: 0.02485:  22%|██▏       | 54/250 [01:53<06:48,  2.08s/it]\n",
      "Loss: 0.02047:  22%|██▏       | 54/250 [01:53<06:48,  2.08s/it]\n",
      "Loss: 0.01861:  22%|██▏       | 54/250 [01:53<06:48,  2.08s/it]\n",
      "Loss: 0.01861:  22%|██▏       | 55/250 [01:53<06:47,  2.09s/it]\n",
      "Loss: 0.02213:  22%|██▏       | 55/250 [01:53<06:47,  2.09s/it]\n",
      "Loss: 0.02151:  22%|██▏       | 55/250 [01:53<06:47,  2.09s/it]\n",
      "Loss: 0.02434:  22%|██▏       | 55/250 [01:53<06:47,  2.09s/it]\n",
      "Loss: 0.01979:  22%|██▏       | 55/250 [01:53<06:47,  2.09s/it]\n",
      "Loss: 0.01823:  22%|██▏       | 55/250 [01:54<06:47,  2.09s/it]\n",
      "Loss: 0.02249:  22%|██▏       | 55/250 [01:54<06:47,  2.09s/it]\n",
      "Loss: 0.02125:  22%|██▏       | 55/250 [01:54<06:47,  2.09s/it]\n",
      "Loss: 0.02073:  22%|██▏       | 55/250 [01:54<06:47,  2.09s/it]\n",
      "Loss: 0.02296:  22%|██▏       | 55/250 [01:54<06:47,  2.09s/it]\n",
      "Loss: 0.02141:  22%|██▏       | 55/250 [01:54<06:47,  2.09s/it]\n",
      "Loss: 0.02474:  22%|██▏       | 55/250 [01:54<06:47,  2.09s/it]\n",
      "Loss: 0.02205:  22%|██▏       | 55/250 [01:54<06:47,  2.09s/it]\n",
      "Loss: 0.02231:  22%|██▏       | 55/250 [01:54<06:47,  2.09s/it]\n",
      "Loss: 0.02385:  22%|██▏       | 55/250 [01:55<06:47,  2.09s/it]\n",
      "Loss: 0.02192:  22%|██▏       | 55/250 [01:55<06:47,  2.09s/it]\n",
      "Loss: 0.02017:  22%|██▏       | 55/250 [01:55<06:47,  2.09s/it]\n",
      "Loss: 0.02366:  22%|██▏       | 55/250 [01:55<06:47,  2.09s/it]\n",
      "Loss: 0.02280:  22%|██▏       | 55/250 [01:55<06:47,  2.09s/it]\n",
      "Loss: 0.01827:  22%|██▏       | 55/250 [01:55<06:47,  2.09s/it]\n",
      "Loss: 0.01827:  22%|██▏       | 56/250 [01:55<06:41,  2.07s/it]\n",
      "Loss: 0.02213:  22%|██▏       | 56/250 [01:55<06:41,  2.07s/it]\n",
      "Loss: 0.02151:  22%|██▏       | 56/250 [01:55<06:41,  2.07s/it]\n",
      "Loss: 0.02627:  22%|██▏       | 56/250 [01:55<06:41,  2.07s/it]\n",
      "Loss: 0.02140:  22%|██▏       | 56/250 [01:55<06:41,  2.07s/it]\n",
      "Loss: 0.02136:  22%|██▏       | 56/250 [01:56<06:41,  2.07s/it]\n",
      "Loss: 0.02095:  22%|██▏       | 56/250 [01:56<06:41,  2.07s/it]\n",
      "Loss: 0.02189:  22%|██▏       | 56/250 [01:56<06:41,  2.07s/it]\n",
      "Loss: 0.02463:  22%|██▏       | 56/250 [01:56<06:41,  2.07s/it]\n",
      "Loss: 0.02248:  22%|██▏       | 56/250 [01:56<06:41,  2.07s/it]\n",
      "Loss: 0.02215:  22%|██▏       | 56/250 [01:56<06:41,  2.07s/it]\n",
      "Loss: 0.02101:  22%|██▏       | 56/250 [01:56<06:41,  2.07s/it]\n",
      "Loss: 0.02572:  22%|██▏       | 56/250 [01:56<06:41,  2.07s/it]\n",
      "Loss: 0.02125:  22%|██▏       | 56/250 [01:56<06:41,  2.07s/it]\n",
      "Loss: 0.02239:  22%|██▏       | 56/250 [01:57<06:41,  2.07s/it]\n",
      "Loss: 0.02050:  22%|██▏       | 56/250 [01:57<06:41,  2.07s/it]\n",
      "Loss: 0.02139:  22%|██▏       | 56/250 [01:57<06:41,  2.07s/it]\n",
      "Loss: 0.02444:  22%|██▏       | 56/250 [01:57<06:41,  2.07s/it]\n",
      "Loss: 0.02143:  22%|██▏       | 56/250 [01:57<06:41,  2.07s/it]\n",
      "Loss: 0.01846:  22%|██▏       | 56/250 [01:57<06:41,  2.07s/it]\n",
      "Loss: 0.01846:  23%|██▎       | 57/250 [01:57<06:39,  2.07s/it]\n",
      "Loss: 0.02225:  23%|██▎       | 57/250 [01:57<06:39,  2.07s/it]\n",
      "Loss: 0.02123:  23%|██▎       | 57/250 [01:57<06:39,  2.07s/it]\n",
      "Loss: 0.02512:  23%|██▎       | 57/250 [01:57<06:39,  2.07s/it]\n",
      "Loss: 0.02153:  23%|██▎       | 57/250 [01:58<06:39,  2.07s/it]\n",
      "Loss: 0.02169:  23%|██▎       | 57/250 [01:58<06:39,  2.07s/it]\n",
      "Loss: 0.01958:  23%|██▎       | 57/250 [01:58<06:39,  2.07s/it]\n",
      "Loss: 0.02132:  23%|██▎       | 57/250 [01:58<06:39,  2.07s/it]\n",
      "Loss: 0.02064:  23%|██▎       | 57/250 [01:58<06:39,  2.07s/it]\n",
      "Loss: 0.02301:  23%|██▎       | 57/250 [01:58<06:39,  2.07s/it]\n",
      "Loss: 0.02092:  23%|██▎       | 57/250 [01:58<06:39,  2.07s/it]\n",
      "Loss: 0.02095:  23%|██▎       | 57/250 [01:58<06:39,  2.07s/it]\n",
      "Loss: 0.02466:  23%|██▎       | 57/250 [01:58<06:39,  2.07s/it]\n",
      "Loss: 0.02141:  23%|██▎       | 57/250 [01:59<06:39,  2.07s/it]\n",
      "Loss: 0.02450:  23%|██▎       | 57/250 [01:59<06:39,  2.07s/it]\n",
      "Loss: 0.01965:  23%|██▎       | 57/250 [01:59<06:39,  2.07s/it]\n",
      "Loss: 0.02056:  23%|██▎       | 57/250 [01:59<06:39,  2.07s/it]\n",
      "Loss: 0.02043:  23%|██▎       | 57/250 [01:59<06:39,  2.07s/it]\n",
      "Loss: 0.02230:  23%|██▎       | 57/250 [01:59<06:39,  2.07s/it]\n",
      "Loss: 0.01837:  23%|██▎       | 57/250 [01:59<06:39,  2.07s/it]\n",
      "Loss: 0.01837:  23%|██▎       | 58/250 [01:59<06:35,  2.06s/it]\n",
      "Loss: 0.02243:  23%|██▎       | 58/250 [01:59<06:35,  2.06s/it]\n",
      "Loss: 0.01954:  23%|██▎       | 58/250 [01:59<06:35,  2.06s/it]\n",
      "Loss: 0.02405:  23%|██▎       | 58/250 [01:59<06:35,  2.06s/it]\n",
      "Loss: 0.02034:  23%|██▎       | 58/250 [02:00<06:35,  2.06s/it]\n",
      "Loss: 0.02150:  23%|██▎       | 58/250 [02:00<06:35,  2.06s/it]\n",
      "Loss: 0.02271:  23%|██▎       | 58/250 [02:00<06:35,  2.06s/it]\n",
      "Loss: 0.02369:  23%|██▎       | 58/250 [02:00<06:35,  2.06s/it]\n",
      "Loss: 0.02354:  23%|██▎       | 58/250 [02:00<06:35,  2.06s/it]\n",
      "Loss: 0.02178:  23%|██▎       | 58/250 [02:00<06:35,  2.06s/it]\n",
      "Loss: 0.02401:  23%|██▎       | 58/250 [02:00<06:35,  2.06s/it]\n",
      "Loss: 0.02212:  23%|██▎       | 58/250 [02:00<06:35,  2.06s/it]\n",
      "Loss: 0.02004:  23%|██▎       | 58/250 [02:00<06:35,  2.06s/it]\n",
      "Loss: 0.02483:  23%|██▎       | 58/250 [02:01<06:35,  2.06s/it]\n",
      "Loss: 0.02208:  23%|██▎       | 58/250 [02:01<06:35,  2.06s/it]\n",
      "Loss: 0.02044:  23%|██▎       | 58/250 [02:01<06:35,  2.06s/it]\n",
      "Loss: 0.02521:  23%|██▎       | 58/250 [02:01<06:35,  2.06s/it]\n",
      "Loss: 0.01976:  23%|██▎       | 58/250 [02:01<06:35,  2.06s/it]\n",
      "Loss: 0.02339:  23%|██▎       | 58/250 [02:01<06:35,  2.06s/it]\n",
      "Loss: 0.01824:  23%|██▎       | 58/250 [02:01<06:35,  2.06s/it]\n",
      "Loss: 0.01824:  24%|██▎       | 59/250 [02:01<06:32,  2.06s/it]\n",
      "Loss: 0.02257:  24%|██▎       | 59/250 [02:01<06:32,  2.06s/it]\n",
      "Loss: 0.02115:  24%|██▎       | 59/250 [02:01<06:32,  2.06s/it]\n",
      "Loss: 0.02372:  24%|██▎       | 59/250 [02:02<06:32,  2.06s/it]\n",
      "Loss: 0.02006:  24%|██▎       | 59/250 [02:02<06:32,  2.06s/it]\n",
      "Loss: 0.01830:  24%|██▎       | 59/250 [02:02<06:32,  2.06s/it]\n",
      "Loss: 0.02232:  24%|██▎       | 59/250 [02:02<06:32,  2.06s/it]\n",
      "Loss: 0.02356:  24%|██▎       | 59/250 [02:02<06:32,  2.06s/it]\n",
      "Loss: 0.02205:  24%|██▎       | 59/250 [02:02<06:32,  2.06s/it]\n",
      "Loss: 0.02326:  24%|██▎       | 59/250 [02:02<06:32,  2.06s/it]\n",
      "Loss: 0.02385:  24%|██▎       | 59/250 [02:02<06:32,  2.06s/it]\n",
      "Loss: 0.02250:  24%|██▎       | 59/250 [02:02<06:32,  2.06s/it]\n",
      "Loss: 0.02253:  24%|██▎       | 59/250 [02:02<06:32,  2.06s/it]\n",
      "Loss: 0.02154:  24%|██▎       | 59/250 [02:03<06:32,  2.06s/it]\n",
      "Loss: 0.02126:  24%|██▎       | 59/250 [02:03<06:32,  2.06s/it]\n",
      "Loss: 0.02056:  24%|██▎       | 59/250 [02:03<06:32,  2.06s/it]\n",
      "Loss: 0.02143:  24%|██▎       | 59/250 [02:03<06:32,  2.06s/it]\n",
      "Loss: 0.02318:  24%|██▎       | 59/250 [02:03<06:32,  2.06s/it]\n",
      "Loss: 0.02093:  24%|██▎       | 59/250 [02:03<06:32,  2.06s/it]\n",
      "Loss: 0.01830:  24%|██▎       | 59/250 [02:03<06:32,  2.06s/it]\n",
      "Loss: 0.01830:  24%|██▍       | 60/250 [02:03<06:31,  2.06s/it]\n",
      "Loss: 0.02242:  24%|██▍       | 60/250 [02:03<06:31,  2.06s/it]\n",
      "Loss: 0.02114:  24%|██▍       | 60/250 [02:03<06:31,  2.06s/it]\n",
      "Loss: 0.02429:  24%|██▍       | 60/250 [02:04<06:31,  2.06s/it]\n",
      "Loss: 0.02120:  24%|██▍       | 60/250 [02:04<06:31,  2.06s/it]\n",
      "Loss: 0.02103:  24%|██▍       | 60/250 [02:04<06:31,  2.06s/it]\n",
      "Loss: 0.02144:  24%|██▍       | 60/250 [02:04<06:31,  2.06s/it]\n",
      "Loss: 0.02188:  24%|██▍       | 60/250 [02:04<06:31,  2.06s/it]\n",
      "Loss: 0.02429:  24%|██▍       | 60/250 [02:04<06:31,  2.06s/it]\n",
      "Loss: 0.02157:  24%|██▍       | 60/250 [02:04<06:31,  2.06s/it]\n",
      "Loss: 0.02310:  24%|██▍       | 60/250 [02:04<06:31,  2.06s/it]\n",
      "Loss: 0.02446:  24%|██▍       | 60/250 [02:04<06:31,  2.06s/it]\n",
      "Loss: 0.02185:  24%|██▍       | 60/250 [02:05<06:31,  2.06s/it]\n",
      "Loss: 0.02232:  24%|██▍       | 60/250 [02:05<06:31,  2.06s/it]\n",
      "Loss: 0.02406:  24%|██▍       | 60/250 [02:05<06:31,  2.06s/it]\n",
      "Loss: 0.02188:  24%|██▍       | 60/250 [02:05<06:31,  2.06s/it]\n",
      "Loss: 0.01991:  24%|██▍       | 60/250 [02:05<06:31,  2.06s/it]\n",
      "Loss: 0.02339:  24%|██▍       | 60/250 [02:05<06:31,  2.06s/it]\n",
      "Loss: 0.02280:  24%|██▍       | 60/250 [02:05<06:31,  2.06s/it]\n",
      "Loss: 0.01841:  24%|██▍       | 60/250 [02:05<06:31,  2.06s/it]\n",
      "Loss: 0.01841:  24%|██▍       | 61/250 [02:05<06:27,  2.05s/it]\n",
      "Loss: 0.02213:  24%|██▍       | 61/250 [02:05<06:27,  2.05s/it]\n",
      "Loss: 0.02139:  24%|██▍       | 61/250 [02:05<06:27,  2.05s/it]\n",
      "Loss: 0.02436:  24%|██▍       | 61/250 [02:06<06:27,  2.05s/it]\n",
      "Loss: 0.02011:  24%|██▍       | 61/250 [02:06<06:27,  2.05s/it]\n",
      "Loss: 0.01814:  24%|██▍       | 61/250 [02:06<06:27,  2.05s/it]\n",
      "Loss: 0.02284:  24%|██▍       | 61/250 [02:06<06:27,  2.05s/it]\n",
      "Loss: 0.02025:  24%|██▍       | 61/250 [02:06<06:27,  2.05s/it]\n",
      "Loss: 0.02076:  24%|██▍       | 61/250 [02:06<06:27,  2.05s/it]\n",
      "Loss: 0.02283:  24%|██▍       | 61/250 [02:06<06:27,  2.05s/it]\n",
      "Loss: 0.02151:  24%|██▍       | 61/250 [02:06<06:27,  2.05s/it]\n",
      "Loss: 0.02427:  24%|██▍       | 61/250 [02:06<06:27,  2.05s/it]\n",
      "Loss: 0.02218:  24%|██▍       | 61/250 [02:07<06:27,  2.05s/it]\n",
      "Loss: 0.02247:  24%|██▍       | 61/250 [02:07<06:27,  2.05s/it]\n",
      "Loss: 0.02505:  24%|██▍       | 61/250 [02:07<06:27,  2.05s/it]\n",
      "Loss: 0.02175:  24%|██▍       | 61/250 [02:07<06:27,  2.05s/it]\n",
      "Loss: 0.01986:  24%|██▍       | 61/250 [02:07<06:27,  2.05s/it]\n",
      "Loss: 0.02315:  24%|██▍       | 61/250 [02:07<06:27,  2.05s/it]\n",
      "Loss: 0.02227:  24%|██▍       | 61/250 [02:07<06:27,  2.05s/it]\n",
      "Loss: 0.01828:  24%|██▍       | 61/250 [02:07<06:27,  2.05s/it]\n",
      "Loss: 0.01828:  25%|██▍       | 62/250 [02:07<06:23,  2.04s/it]\n",
      "Loss: 0.02234:  25%|██▍       | 62/250 [02:07<06:23,  2.04s/it]\n",
      "Loss: 0.02111:  25%|██▍       | 62/250 [02:08<06:23,  2.04s/it]\n",
      "Loss: 0.02455:  25%|██▍       | 62/250 [02:08<06:23,  2.04s/it]\n",
      "Loss: 0.02127:  25%|██▍       | 62/250 [02:08<06:23,  2.04s/it]\n",
      "Loss: 0.02143:  25%|██▍       | 62/250 [02:08<06:23,  2.04s/it]\n",
      "Loss: 0.02149:  25%|██▍       | 62/250 [02:08<06:23,  2.04s/it]\n",
      "Loss: 0.02171:  25%|██▍       | 62/250 [02:08<06:23,  2.04s/it]\n",
      "Loss: 0.02255:  25%|██▍       | 62/250 [02:08<06:23,  2.04s/it]\n",
      "Loss: 0.01931:  25%|██▍       | 62/250 [02:08<06:23,  2.04s/it]\n",
      "Loss: 0.02397:  25%|██▍       | 62/250 [02:08<06:23,  2.04s/it]\n",
      "Loss: 0.02438:  25%|██▍       | 62/250 [02:08<06:23,  2.04s/it]\n",
      "Loss: 0.02242:  25%|██▍       | 62/250 [02:09<06:23,  2.04s/it]\n",
      "Loss: 0.02211:  25%|██▍       | 62/250 [02:09<06:23,  2.04s/it]\n",
      "Loss: 0.02088:  25%|██▍       | 62/250 [02:09<06:23,  2.04s/it]\n",
      "Loss: 0.02146:  25%|██▍       | 62/250 [02:09<06:23,  2.04s/it]\n",
      "Loss: 0.02329:  25%|██▍       | 62/250 [02:09<06:23,  2.04s/it]\n",
      "Loss: 0.02133:  25%|██▍       | 62/250 [02:09<06:23,  2.04s/it]\n",
      "Loss: 0.01842:  25%|██▍       | 62/250 [02:09<06:23,  2.04s/it]\n",
      "Loss: 0.01842:  25%|██▌       | 63/250 [02:09<06:18,  2.02s/it]\n",
      "Loss: 0.02229:  25%|██▌       | 63/250 [02:09<06:18,  2.02s/it]\n",
      "Loss: 0.02104:  25%|██▌       | 63/250 [02:09<06:18,  2.02s/it]\n",
      "Loss: 0.02493:  25%|██▌       | 63/250 [02:10<06:18,  2.02s/it]\n",
      "Loss: 0.02007:  25%|██▌       | 63/250 [02:10<06:18,  2.02s/it]\n",
      "Loss: 0.02137:  25%|██▌       | 63/250 [02:10<06:18,  2.02s/it]\n",
      "Loss: 0.02277:  25%|██▌       | 63/250 [02:10<06:18,  2.02s/it]\n",
      "Loss: 0.02025:  25%|██▌       | 63/250 [02:10<06:18,  2.02s/it]\n",
      "Loss: 0.02227:  25%|██▌       | 63/250 [02:10<06:18,  2.02s/it]\n",
      "Loss: 0.02288:  25%|██▌       | 63/250 [02:10<06:18,  2.02s/it]\n",
      "Loss: 0.02224:  25%|██▌       | 63/250 [02:10<06:18,  2.02s/it]\n",
      "Loss: 0.02026:  25%|██▌       | 63/250 [02:10<06:18,  2.02s/it]\n",
      "Loss: 0.02114:  25%|██▌       | 63/250 [02:11<06:18,  2.02s/it]\n",
      "Loss: 0.02187:  25%|██▌       | 63/250 [02:11<06:18,  2.02s/it]\n",
      "Loss: 0.01946:  25%|██▌       | 63/250 [02:11<06:18,  2.02s/it]\n",
      "Loss: 0.02353:  25%|██▌       | 63/250 [02:11<06:18,  2.02s/it]\n",
      "Loss: 0.02546:  25%|██▌       | 63/250 [02:11<06:18,  2.02s/it]\n",
      "Loss: 0.02451:  25%|██▌       | 63/250 [02:11<06:18,  2.02s/it]\n",
      "Loss: 0.02061:  25%|██▌       | 63/250 [02:11<06:18,  2.02s/it]\n",
      "Loss: 0.01839:  25%|██▌       | 63/250 [02:11<06:18,  2.02s/it]\n",
      "Loss: 0.01839:  26%|██▌       | 64/250 [02:11<06:16,  2.02s/it]\n",
      "Loss: 0.02215:  26%|██▌       | 64/250 [02:11<06:16,  2.02s/it]\n",
      "Loss: 0.01936:  26%|██▌       | 64/250 [02:12<06:16,  2.02s/it]\n",
      "Loss: 0.02380:  26%|██▌       | 64/250 [02:12<06:16,  2.02s/it]\n",
      "Loss: 0.02011:  26%|██▌       | 64/250 [02:12<06:16,  2.02s/it]\n",
      "Loss: 0.02125:  26%|██▌       | 64/250 [02:12<06:16,  2.02s/it]\n",
      "Loss: 0.02224:  26%|██▌       | 64/250 [02:12<06:16,  2.02s/it]\n",
      "Loss: 0.02325:  26%|██▌       | 64/250 [02:12<06:16,  2.02s/it]\n",
      "Loss: 0.02223:  26%|██▌       | 64/250 [02:12<06:16,  2.02s/it]\n",
      "Loss: 0.02319:  26%|██▌       | 64/250 [02:12<06:16,  2.02s/it]\n",
      "Loss: 0.02202:  26%|██▌       | 64/250 [02:12<06:16,  2.02s/it]\n",
      "Loss: 0.02117:  26%|██▌       | 64/250 [02:12<06:16,  2.02s/it]\n",
      "Loss: 0.02186:  26%|██▌       | 64/250 [02:13<06:16,  2.02s/it]\n",
      "Loss: 0.02190:  26%|██▌       | 64/250 [02:13<06:16,  2.02s/it]\n",
      "Loss: 0.01916:  26%|██▌       | 64/250 [02:13<06:16,  2.02s/it]\n",
      "Loss: 0.02372:  26%|██▌       | 64/250 [02:13<06:16,  2.02s/it]\n",
      "Loss: 0.02506:  26%|██▌       | 64/250 [02:13<06:16,  2.02s/it]\n",
      "Loss: 0.02440:  26%|██▌       | 64/250 [02:13<06:16,  2.02s/it]\n",
      "Loss: 0.02008:  26%|██▌       | 64/250 [02:13<06:16,  2.02s/it]\n",
      "Loss: 0.01857:  26%|██▌       | 64/250 [02:13<06:16,  2.02s/it]\n",
      "Loss: 0.01857:  26%|██▌       | 65/250 [02:13<06:15,  2.03s/it]\n",
      "Loss: 0.02250:  26%|██▌       | 65/250 [02:13<06:15,  2.03s/it]\n",
      "Loss: 0.02130:  26%|██▌       | 65/250 [02:14<06:15,  2.03s/it]\n",
      "Loss: 0.02426:  26%|██▌       | 65/250 [02:14<06:15,  2.03s/it]\n",
      "Loss: 0.01995:  26%|██▌       | 65/250 [02:14<06:15,  2.03s/it]\n",
      "Loss: 0.01828:  26%|██▌       | 65/250 [02:14<06:15,  2.03s/it]\n",
      "Loss: 0.02241:  26%|██▌       | 65/250 [02:14<06:15,  2.03s/it]\n",
      "Loss: 0.02046:  26%|██▌       | 65/250 [02:14<06:15,  2.03s/it]\n",
      "Loss: 0.02228:  26%|██▌       | 65/250 [02:14<06:15,  2.03s/it]\n",
      "Loss: 0.02312:  26%|██▌       | 65/250 [02:14<06:15,  2.03s/it]\n",
      "Loss: 0.02080:  26%|██▌       | 65/250 [02:14<06:15,  2.03s/it]\n",
      "Loss: 0.02093:  26%|██▌       | 65/250 [02:15<06:15,  2.03s/it]\n",
      "Loss: 0.02449:  26%|██▌       | 65/250 [02:15<06:15,  2.03s/it]\n",
      "Loss: 0.02142:  26%|██▌       | 65/250 [02:15<06:15,  2.03s/it]\n",
      "Loss: 0.02430:  26%|██▌       | 65/250 [02:15<06:15,  2.03s/it]\n",
      "Loss: 0.02024:  26%|██▌       | 65/250 [02:15<06:15,  2.03s/it]\n",
      "Loss: 0.02028:  26%|██▌       | 65/250 [02:15<06:15,  2.03s/it]\n",
      "Loss: 0.02041:  26%|██▌       | 65/250 [02:15<06:15,  2.03s/it]\n",
      "Loss: 0.02175:  26%|██▌       | 65/250 [02:15<06:15,  2.03s/it]\n",
      "Loss: 0.01854:  26%|██▌       | 65/250 [02:15<06:15,  2.03s/it]\n",
      "Loss: 0.01854:  26%|██▋       | 66/250 [02:15<06:13,  2.03s/it]\n",
      "Loss: 0.02232:  26%|██▋       | 66/250 [02:15<06:13,  2.03s/it]\n",
      "Loss: 0.01923:  26%|██▋       | 66/250 [02:16<06:13,  2.03s/it]\n",
      "Loss: 0.02389:  26%|██▋       | 66/250 [02:16<06:13,  2.03s/it]\n",
      "Loss: 0.02005:  26%|██▋       | 66/250 [02:16<06:13,  2.03s/it]\n",
      "Loss: 0.02130:  26%|██▋       | 66/250 [02:16<06:13,  2.03s/it]\n",
      "Loss: 0.02026:  26%|██▋       | 66/250 [02:16<06:13,  2.03s/it]\n",
      "Loss: 0.02030:  26%|██▋       | 66/250 [02:16<06:13,  2.03s/it]\n",
      "Loss: 0.02107:  26%|██▋       | 66/250 [02:16<06:13,  2.03s/it]\n",
      "Loss: 0.02264:  26%|██▋       | 66/250 [02:16<06:13,  2.03s/it]\n",
      "Loss: 0.02170:  26%|██▋       | 66/250 [02:16<06:13,  2.03s/it]\n",
      "Loss: 0.02045:  26%|██▋       | 66/250 [02:17<06:13,  2.03s/it]\n",
      "Loss: 0.02112:  26%|██▋       | 66/250 [02:17<06:13,  2.03s/it]\n",
      "Loss: 0.02192:  26%|██▋       | 66/250 [02:17<06:13,  2.03s/it]\n",
      "Loss: 0.01908:  26%|██▋       | 66/250 [02:17<06:13,  2.03s/it]\n",
      "Loss: 0.02358:  26%|██▋       | 66/250 [02:17<06:13,  2.03s/it]\n",
      "Loss: 0.02501:  26%|██▋       | 66/250 [02:17<06:13,  2.03s/it]\n",
      "Loss: 0.02442:  26%|██▋       | 66/250 [02:17<06:13,  2.03s/it]\n",
      "Loss: 0.01997:  26%|██▋       | 66/250 [02:17<06:13,  2.03s/it]\n",
      "Loss: 0.01847:  26%|██▋       | 66/250 [02:17<06:13,  2.03s/it]\n",
      "Loss: 0.01847:  27%|██▋       | 67/250 [02:17<06:13,  2.04s/it]\n",
      "Loss: 0.02243:  27%|██▋       | 67/250 [02:18<06:13,  2.04s/it]\n",
      "Loss: 0.02117:  27%|██▋       | 67/250 [02:18<06:13,  2.04s/it]\n",
      "Loss: 0.02435:  27%|██▋       | 67/250 [02:18<06:13,  2.04s/it]\n",
      "Loss: 0.02128:  27%|██▋       | 67/250 [02:18<06:13,  2.04s/it]\n",
      "Loss: 0.02168:  27%|██▋       | 67/250 [02:18<06:13,  2.04s/it]\n",
      "Loss: 0.02271:  27%|██▋       | 67/250 [02:18<06:13,  2.04s/it]\n",
      "Loss: 0.02025:  27%|██▋       | 67/250 [02:18<06:13,  2.04s/it]\n",
      "Loss: 0.02202:  27%|██▋       | 67/250 [02:18<06:13,  2.04s/it]\n",
      "Loss: 0.02242:  27%|██▋       | 67/250 [02:18<06:13,  2.04s/it]\n",
      "Loss: 0.02191:  27%|██▋       | 67/250 [02:19<06:13,  2.04s/it]\n",
      "Loss: 0.02033:  27%|██▋       | 67/250 [02:19<06:13,  2.04s/it]\n",
      "Loss: 0.02113:  27%|██▋       | 67/250 [02:19<06:13,  2.04s/it]\n",
      "Loss: 0.02182:  27%|██▋       | 67/250 [02:19<06:13,  2.04s/it]\n",
      "Loss: 0.01919:  27%|██▋       | 67/250 [02:19<06:13,  2.04s/it]\n",
      "Loss: 0.02342:  27%|██▋       | 67/250 [02:19<06:13,  2.04s/it]\n",
      "Loss: 0.02528:  27%|██▋       | 67/250 [02:19<06:13,  2.04s/it]\n",
      "Loss: 0.02464:  27%|██▋       | 67/250 [02:19<06:13,  2.04s/it]\n",
      "Loss: 0.02033:  27%|██▋       | 67/250 [02:19<06:13,  2.04s/it]\n",
      "Loss: 0.01824:  27%|██▋       | 67/250 [02:19<06:13,  2.04s/it]\n",
      "Loss: 0.01824:  27%|██▋       | 68/250 [02:19<06:11,  2.04s/it]\n",
      "Loss: 0.02227:  27%|██▋       | 68/250 [02:20<06:11,  2.04s/it]\n",
      "Loss: 0.02117:  27%|██▋       | 68/250 [02:20<06:11,  2.04s/it]\n",
      "Loss: 0.02417:  27%|██▋       | 68/250 [02:20<06:11,  2.04s/it]\n",
      "Loss: 0.02139:  27%|██▋       | 68/250 [02:20<06:11,  2.04s/it]\n",
      "Loss: 0.02075:  27%|██▋       | 68/250 [02:20<06:11,  2.04s/it]\n",
      "Loss: 0.02098:  27%|██▋       | 68/250 [02:20<06:11,  2.04s/it]\n",
      "Loss: 0.02171:  27%|██▋       | 68/250 [02:20<06:11,  2.04s/it]\n",
      "Loss: 0.02204:  27%|██▋       | 68/250 [02:20<06:11,  2.04s/it]\n",
      "Loss: 0.01910:  27%|██▋       | 68/250 [02:20<06:11,  2.04s/it]\n",
      "Loss: 0.02343:  27%|██▋       | 68/250 [02:21<06:11,  2.04s/it]\n",
      "Loss: 0.02416:  27%|██▋       | 68/250 [02:21<06:11,  2.04s/it]\n",
      "Loss: 0.02254:  27%|██▋       | 68/250 [02:21<06:11,  2.04s/it]\n",
      "Loss: 0.02198:  27%|██▋       | 68/250 [02:21<06:11,  2.04s/it]\n",
      "Loss: 0.02054:  27%|██▋       | 68/250 [02:21<06:11,  2.04s/it]\n",
      "Loss: 0.02167:  27%|██▋       | 68/250 [02:21<06:11,  2.04s/it]\n",
      "Loss: 0.02309:  27%|██▋       | 68/250 [02:21<06:11,  2.04s/it]\n",
      "Loss: 0.02126:  27%|██▋       | 68/250 [02:21<06:11,  2.04s/it]\n",
      "Loss: 0.01828:  27%|██▋       | 68/250 [02:21<06:11,  2.04s/it]\n",
      "Loss: 0.01828:  28%|██▊       | 69/250 [02:21<06:06,  2.03s/it]\n",
      "Loss: 0.02236:  28%|██▊       | 69/250 [02:22<06:06,  2.03s/it]\n",
      "Loss: 0.02116:  28%|██▊       | 69/250 [02:22<06:06,  2.03s/it]\n",
      "Loss: 0.02358:  28%|██▊       | 69/250 [02:22<06:06,  2.03s/it]\n",
      "Loss: 0.01982:  28%|██▊       | 69/250 [02:22<06:06,  2.03s/it]\n",
      "Loss: 0.02127:  28%|██▊       | 69/250 [02:22<06:06,  2.03s/it]\n",
      "Loss: 0.01951:  28%|██▊       | 69/250 [02:22<06:06,  2.03s/it]\n",
      "Loss: 0.02112:  28%|██▊       | 69/250 [02:22<06:06,  2.03s/it]\n",
      "Loss: 0.02414:  28%|██▊       | 69/250 [02:22<06:06,  2.03s/it]\n",
      "Loss: 0.02244:  28%|██▊       | 69/250 [02:22<06:06,  2.03s/it]\n",
      "Loss: 0.02142:  28%|██▊       | 69/250 [02:23<06:06,  2.03s/it]\n",
      "Loss: 0.02014:  28%|██▊       | 69/250 [02:23<06:06,  2.03s/it]\n",
      "Loss: 0.02441:  28%|██▊       | 69/250 [02:23<06:06,  2.03s/it]\n",
      "Loss: 0.02467:  28%|██▊       | 69/250 [02:23<06:06,  2.03s/it]\n",
      "Loss: 0.02199:  28%|██▊       | 69/250 [02:23<06:06,  2.03s/it]\n",
      "Loss: 0.02048:  28%|██▊       | 69/250 [02:23<06:06,  2.03s/it]\n",
      "Loss: 0.02506:  28%|██▊       | 69/250 [02:23<06:06,  2.03s/it]\n",
      "Loss: 0.01997:  28%|██▊       | 69/250 [02:23<06:06,  2.03s/it]\n",
      "Loss: 0.02161:  28%|██▊       | 69/250 [02:23<06:06,  2.03s/it]\n",
      "Loss: 0.01889:  28%|██▊       | 69/250 [02:23<06:06,  2.03s/it]\n",
      "Loss: 0.01889:  28%|██▊       | 70/250 [02:24<06:05,  2.03s/it]\n",
      "Loss: 0.02204:  28%|██▊       | 70/250 [02:24<06:05,  2.03s/it]\n",
      "Loss: 0.02113:  28%|██▊       | 70/250 [02:24<06:05,  2.03s/it]\n",
      "Loss: 0.02394:  28%|██▊       | 70/250 [02:24<06:05,  2.03s/it]\n",
      "Loss: 0.02008:  28%|██▊       | 70/250 [02:24<06:05,  2.03s/it]\n",
      "Loss: 0.01816:  28%|██▊       | 70/250 [02:24<06:05,  2.03s/it]\n",
      "Loss: 0.02241:  28%|██▊       | 70/250 [02:24<06:05,  2.03s/it]\n",
      "Loss: 0.02027:  28%|██▊       | 70/250 [02:24<06:05,  2.03s/it]\n",
      "Loss: 0.02020:  28%|██▊       | 70/250 [02:24<06:05,  2.03s/it]\n",
      "Loss: 0.02237:  28%|██▊       | 70/250 [02:24<06:05,  2.03s/it]\n",
      "Loss: 0.02145:  28%|██▊       | 70/250 [02:25<06:05,  2.03s/it]\n",
      "Loss: 0.02439:  28%|██▊       | 70/250 [02:25<06:05,  2.03s/it]\n",
      "Loss: 0.02173:  28%|██▊       | 70/250 [02:25<06:05,  2.03s/it]\n",
      "Loss: 0.02221:  28%|██▊       | 70/250 [02:25<06:05,  2.03s/it]\n",
      "Loss: 0.02389:  28%|██▊       | 70/250 [02:25<06:05,  2.03s/it]\n",
      "Loss: 0.02211:  28%|██▊       | 70/250 [02:25<06:05,  2.03s/it]\n",
      "Loss: 0.01986:  28%|██▊       | 70/250 [02:25<06:05,  2.03s/it]\n",
      "Loss: 0.02315:  28%|██▊       | 70/250 [02:25<06:05,  2.03s/it]\n",
      "Loss: 0.02248:  28%|██▊       | 70/250 [02:26<06:05,  2.03s/it]\n",
      "Loss: 0.01823:  28%|██▊       | 70/250 [02:26<06:05,  2.03s/it]\n",
      "Loss: 0.01823:  28%|██▊       | 71/250 [02:26<06:03,  2.03s/it]\n",
      "Loss: 0.02202:  28%|██▊       | 71/250 [02:26<06:03,  2.03s/it]\n",
      "Loss: 0.02133:  28%|██▊       | 71/250 [02:26<06:03,  2.03s/it]\n",
      "Loss: 0.02392:  28%|██▊       | 71/250 [02:26<06:03,  2.03s/it]\n",
      "Loss: 0.02125:  28%|██▊       | 71/250 [02:26<06:03,  2.03s/it]\n",
      "Loss: 0.02089:  28%|██▊       | 71/250 [02:26<06:03,  2.03s/it]\n",
      "Loss: 0.02090:  28%|██▊       | 71/250 [02:26<06:03,  2.03s/it]\n",
      "Loss: 0.02186:  28%|██▊       | 71/250 [02:26<06:03,  2.03s/it]\n",
      "Loss: 0.02213:  28%|██▊       | 71/250 [02:26<06:03,  2.03s/it]\n",
      "Loss: 0.01937:  28%|██▊       | 71/250 [02:27<06:03,  2.03s/it]\n",
      "Loss: 0.02325:  28%|██▊       | 71/250 [02:27<06:03,  2.03s/it]\n",
      "Loss: 0.02482:  28%|██▊       | 71/250 [02:27<06:03,  2.03s/it]\n",
      "Loss: 0.02222:  28%|██▊       | 71/250 [02:27<06:03,  2.03s/it]\n",
      "Loss: 0.02200:  28%|██▊       | 71/250 [02:27<06:03,  2.03s/it]\n",
      "Loss: 0.02089:  28%|██▊       | 71/250 [02:27<06:03,  2.03s/it]\n",
      "Loss: 0.02134:  28%|██▊       | 71/250 [02:27<06:03,  2.03s/it]\n",
      "Loss: 0.02295:  28%|██▊       | 71/250 [02:27<06:03,  2.03s/it]\n",
      "Loss: 0.02137:  28%|██▊       | 71/250 [02:27<06:03,  2.03s/it]\n",
      "Loss: 0.01839:  28%|██▊       | 71/250 [02:28<06:03,  2.03s/it]\n",
      "Loss: 0.01839:  29%|██▉       | 72/250 [02:28<06:00,  2.03s/it]\n",
      "Loss: 0.02245:  29%|██▉       | 72/250 [02:28<06:00,  2.03s/it]\n",
      "Loss: 0.02122:  29%|██▉       | 72/250 [02:28<06:00,  2.03s/it]\n",
      "Loss: 0.02374:  29%|██▉       | 72/250 [02:28<06:00,  2.03s/it]\n",
      "Loss: 0.01993:  29%|██▉       | 72/250 [02:28<06:00,  2.03s/it]\n",
      "Loss: 0.02117:  29%|██▉       | 72/250 [02:28<06:00,  2.03s/it]\n",
      "Loss: 0.01951:  29%|██▉       | 72/250 [02:28<06:00,  2.03s/it]\n",
      "Loss: 0.02136:  29%|██▉       | 72/250 [02:28<06:00,  2.03s/it]\n",
      "Loss: 0.02442:  29%|██▉       | 72/250 [02:28<06:00,  2.03s/it]\n",
      "Loss: 0.02204:  29%|██▉       | 72/250 [02:29<06:00,  2.03s/it]\n",
      "Loss: 0.02332:  29%|██▉       | 72/250 [02:29<06:00,  2.03s/it]\n",
      "Loss: 0.02198:  29%|██▉       | 72/250 [02:29<06:00,  2.03s/it]\n",
      "Loss: 0.02381:  29%|██▉       | 72/250 [02:29<06:00,  2.03s/it]\n",
      "Loss: 0.02059:  29%|██▉       | 72/250 [02:29<06:00,  2.03s/it]\n",
      "Loss: 0.02135:  29%|██▉       | 72/250 [02:29<06:00,  2.03s/it]\n",
      "Loss: 0.01965:  29%|██▉       | 72/250 [02:29<06:00,  2.03s/it]\n",
      "Loss: 0.01948:  29%|██▉       | 72/250 [02:29<06:00,  2.03s/it]\n",
      "Loss: 0.02266:  29%|██▉       | 72/250 [02:29<06:00,  2.03s/it]\n",
      "Loss: 0.01813:  29%|██▉       | 72/250 [02:30<06:00,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.39it/s]\u001b[A\n",
      "Loss: 0.01813:  29%|██▉       | 73/250 [02:30<05:58,  2.02s/it]\n",
      "Loss: 0.02075:  29%|██▉       | 73/250 [02:30<05:58,  2.02s/it]\n",
      "Loss: 0.01972:  29%|██▉       | 73/250 [02:30<05:58,  2.02s/it]\n",
      "Loss: 0.01961:  29%|██▉       | 73/250 [02:30<05:58,  2.02s/it]\n",
      "Loss: 0.02108:  29%|██▉       | 73/250 [02:30<05:58,  2.02s/it]\n",
      "Loss: 0.02283:  29%|██▉       | 73/250 [02:30<05:58,  2.02s/it]\n",
      "Loss: 0.02099:  29%|██▉       | 73/250 [02:30<05:58,  2.02s/it]\n",
      "Loss: 0.02110:  29%|██▉       | 73/250 [02:30<05:58,  2.02s/it]\n",
      "Loss: 0.02452:  29%|██▉       | 73/250 [02:30<05:58,  2.02s/it]\n",
      "Loss: 0.02200:  29%|██▉       | 73/250 [02:31<05:58,  2.02s/it]\n",
      "Loss: 0.02228:  29%|██▉       | 73/250 [02:31<05:58,  2.02s/it]\n",
      "Loss: 0.02248:  29%|██▉       | 73/250 [02:31<05:58,  2.02s/it]\n",
      "Loss: 0.02246:  29%|██▉       | 73/250 [02:31<05:58,  2.02s/it]\n",
      "Loss: 0.02123:  29%|██▉       | 73/250 [02:31<05:58,  2.02s/it]\n",
      "Loss: 0.02112:  29%|██▉       | 73/250 [02:31<05:58,  2.02s/it]\n",
      "Loss: 0.02024:  29%|██▉       | 73/250 [02:31<05:58,  2.02s/it]\n",
      "Loss: 0.02146:  29%|██▉       | 73/250 [02:31<05:58,  2.02s/it]\n",
      "Loss: 0.02313:  29%|██▉       | 73/250 [02:31<05:58,  2.02s/it]\n",
      "Loss: 0.02119:  29%|██▉       | 73/250 [02:32<05:58,  2.02s/it]\n",
      "Loss: 0.01817:  29%|██▉       | 73/250 [02:32<05:58,  2.02s/it]\n",
      "Loss: 0.01817:  30%|██▉       | 74/250 [02:32<05:58,  2.03s/it]\n",
      "Loss: 0.02199:  30%|██▉       | 74/250 [02:32<05:58,  2.03s/it]\n",
      "Loss: 0.02094:  30%|██▉       | 74/250 [02:32<05:58,  2.03s/it]\n",
      "Loss: 0.02405:  30%|██▉       | 74/250 [02:32<05:58,  2.03s/it]\n",
      "Loss: 0.02148:  30%|██▉       | 74/250 [02:32<05:58,  2.03s/it]\n",
      "Loss: 0.02046:  30%|██▉       | 74/250 [02:32<05:58,  2.03s/it]\n",
      "Loss: 0.01932:  30%|██▉       | 74/250 [02:32<05:58,  2.03s/it]\n",
      "Loss: 0.02112:  30%|██▉       | 74/250 [02:32<05:58,  2.03s/it]\n",
      "Loss: 0.02441:  30%|██▉       | 74/250 [02:33<05:58,  2.03s/it]\n",
      "Loss: 0.02132:  30%|██▉       | 74/250 [02:33<05:58,  2.03s/it]\n",
      "Loss: 0.02297:  30%|██▉       | 74/250 [02:33<05:58,  2.03s/it]\n",
      "Loss: 0.02203:  30%|██▉       | 74/250 [02:33<05:58,  2.03s/it]\n",
      "Loss: 0.02384:  30%|██▉       | 74/250 [02:33<05:58,  2.03s/it]\n",
      "Loss: 0.02062:  30%|██▉       | 74/250 [02:33<05:58,  2.03s/it]\n",
      "Loss: 0.02173:  30%|██▉       | 74/250 [02:33<05:58,  2.03s/it]\n",
      "Loss: 0.02007:  30%|██▉       | 74/250 [02:33<05:58,  2.03s/it]\n",
      "Loss: 0.01981:  30%|██▉       | 74/250 [02:33<05:58,  2.03s/it]\n",
      "Loss: 0.02242:  30%|██▉       | 74/250 [02:34<05:58,  2.03s/it]\n",
      "Loss: 0.01823:  30%|██▉       | 74/250 [02:34<05:58,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.71it/s]\u001b[A\n",
      "Loss: 0.01823:  30%|███       | 75/250 [02:34<05:56,  2.03s/it]\n",
      "Loss: 0.02201:  30%|███       | 75/250 [02:34<05:56,  2.03s/it]\n",
      "Loss: 0.02118:  30%|███       | 75/250 [02:34<05:56,  2.03s/it]\n",
      "Loss: 0.02383:  30%|███       | 75/250 [02:34<05:56,  2.03s/it]\n",
      "Loss: 0.02130:  30%|███       | 75/250 [02:34<05:56,  2.03s/it]\n",
      "Loss: 0.02053:  30%|███       | 75/250 [02:34<05:56,  2.03s/it]\n",
      "Loss: 0.01948:  30%|███       | 75/250 [02:34<05:56,  2.03s/it]\n",
      "Loss: 0.02130:  30%|███       | 75/250 [02:34<05:56,  2.03s/it]\n",
      "Loss: 0.02395:  30%|███       | 75/250 [02:35<05:56,  2.03s/it]\n",
      "Loss: 0.02260:  30%|███       | 75/250 [02:35<05:56,  2.03s/it]\n",
      "Loss: 0.02196:  30%|███       | 75/250 [02:35<05:56,  2.03s/it]\n",
      "Loss: 0.02490:  30%|███       | 75/250 [02:35<05:56,  2.03s/it]\n",
      "Loss: 0.02169:  30%|███       | 75/250 [02:35<05:56,  2.03s/it]\n",
      "Loss: 0.02221:  30%|███       | 75/250 [02:35<05:56,  2.03s/it]\n",
      "Loss: 0.02377:  30%|███       | 75/250 [02:35<05:56,  2.03s/it]\n",
      "Loss: 0.02146:  30%|███       | 75/250 [02:35<05:56,  2.03s/it]\n",
      "Loss: 0.01980:  30%|███       | 75/250 [02:35<05:56,  2.03s/it]\n",
      "Loss: 0.02310:  30%|███       | 75/250 [02:36<05:56,  2.03s/it]\n",
      "Loss: 0.02267:  30%|███       | 75/250 [02:36<05:56,  2.03s/it]\n",
      "Loss: 0.01838:  30%|███       | 75/250 [02:36<05:56,  2.03s/it]\n",
      "Loss: 0.01838:  30%|███       | 76/250 [02:36<05:54,  2.04s/it]\n",
      "Loss: 0.02201:  30%|███       | 76/250 [02:36<05:54,  2.04s/it]\n",
      "Loss: 0.02174:  30%|███       | 76/250 [02:36<05:54,  2.04s/it]\n",
      "Loss: 0.02048:  30%|███       | 76/250 [02:36<05:54,  2.04s/it]\n",
      "Loss: 0.02307:  30%|███       | 76/250 [02:36<05:54,  2.04s/it]\n",
      "Loss: 0.02139:  30%|███       | 76/250 [02:36<05:54,  2.04s/it]\n",
      "Loss: 0.02088:  30%|███       | 76/250 [02:36<05:54,  2.04s/it]\n",
      "Loss: 0.02207:  30%|███       | 76/250 [02:36<05:54,  2.04s/it]\n",
      "Loss: 0.02160:  30%|███       | 76/250 [02:37<05:54,  2.04s/it]\n",
      "Loss: 0.01918:  30%|███       | 76/250 [02:37<05:54,  2.04s/it]\n",
      "Loss: 0.02318:  30%|███       | 76/250 [02:37<05:54,  2.04s/it]\n",
      "Loss: 0.02400:  30%|███       | 76/250 [02:37<05:54,  2.04s/it]\n",
      "Loss: 0.02208:  30%|███       | 76/250 [02:37<05:54,  2.04s/it]\n",
      "Loss: 0.02193:  30%|███       | 76/250 [02:37<05:54,  2.04s/it]\n",
      "Loss: 0.02030:  30%|███       | 76/250 [02:37<05:54,  2.04s/it]\n",
      "Loss: 0.02134:  30%|███       | 76/250 [02:37<05:54,  2.04s/it]\n",
      "Loss: 0.02322:  30%|███       | 76/250 [02:38<05:54,  2.04s/it]\n",
      "Loss: 0.02140:  30%|███       | 76/250 [02:38<05:54,  2.04s/it]\n",
      "Loss: 0.01824:  30%|███       | 76/250 [02:38<05:54,  2.04s/it]\n",
      "Loss: 0.01824:  31%|███       | 77/250 [02:38<05:51,  2.03s/it]\n",
      "Loss: 0.02075:  31%|███       | 77/250 [02:38<05:51,  2.03s/it]\n",
      "Loss: 0.01922:  31%|███       | 77/250 [02:38<05:51,  2.03s/it]\n",
      "Loss: 0.02345:  31%|███       | 77/250 [02:38<05:51,  2.03s/it]\n",
      "Loss: 0.01964:  31%|███       | 77/250 [02:38<05:51,  2.03s/it]\n",
      "Loss: 0.02220:  31%|███       | 77/250 [02:38<05:51,  2.03s/it]\n",
      "Loss: 0.02022:  31%|███       | 77/250 [02:38<05:51,  2.03s/it]\n",
      "Loss: 0.02104:  31%|███       | 77/250 [02:38<05:51,  2.03s/it]\n",
      "Loss: 0.02423:  31%|███       | 77/250 [02:39<05:51,  2.03s/it]\n",
      "Loss: 0.02131:  31%|███       | 77/250 [02:39<05:51,  2.03s/it]\n",
      "Loss: 0.02308:  31%|███       | 77/250 [02:39<05:51,  2.03s/it]\n",
      "Loss: 0.02237:  31%|███       | 77/250 [02:39<05:51,  2.03s/it]\n",
      "Loss: 0.02377:  31%|███       | 77/250 [02:39<05:51,  2.03s/it]\n",
      "Loss: 0.02072:  31%|███       | 77/250 [02:39<05:51,  2.03s/it]\n",
      "Loss: 0.02159:  31%|███       | 77/250 [02:39<05:51,  2.03s/it]\n",
      "Loss: 0.01999:  31%|███       | 77/250 [02:39<05:51,  2.03s/it]\n",
      "Loss: 0.01952:  31%|███       | 77/250 [02:40<05:51,  2.03s/it]\n",
      "Loss: 0.02187:  31%|███       | 77/250 [02:40<05:51,  2.03s/it]\n",
      "Loss: 0.01824:  31%|███       | 77/250 [02:40<05:51,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 62.48it/s]\u001b[A\n",
      "Loss: 0.01824:  31%|███       | 78/250 [02:40<05:48,  2.03s/it]\n",
      "Loss: 0.02207:  31%|███       | 78/250 [02:40<05:48,  2.03s/it]\n",
      "Loss: 0.02093:  31%|███       | 78/250 [02:40<05:48,  2.03s/it]\n",
      "Loss: 0.02519:  31%|███       | 78/250 [02:40<05:48,  2.03s/it]\n",
      "Loss: 0.02119:  31%|███       | 78/250 [02:40<05:48,  2.03s/it]\n",
      "Loss: 0.02075:  31%|███       | 78/250 [02:40<05:48,  2.03s/it]\n",
      "Loss: 0.02104:  31%|███       | 78/250 [02:40<05:48,  2.03s/it]\n",
      "Loss: 0.02160:  31%|███       | 78/250 [02:41<05:48,  2.03s/it]\n",
      "Loss: 0.02415:  31%|███       | 78/250 [02:41<05:48,  2.03s/it]\n",
      "Loss: 0.02242:  31%|███       | 78/250 [02:41<05:48,  2.03s/it]\n",
      "Loss: 0.02193:  31%|███       | 78/250 [02:41<05:48,  2.03s/it]\n",
      "Loss: 0.02029:  31%|███       | 78/250 [02:41<05:48,  2.03s/it]\n",
      "Loss: 0.02454:  31%|███       | 78/250 [02:41<05:48,  2.03s/it]\n",
      "Loss: 0.02118:  31%|███       | 78/250 [02:41<05:48,  2.03s/it]\n",
      "Loss: 0.02383:  31%|███       | 78/250 [02:41<05:48,  2.03s/it]\n",
      "Loss: 0.01975:  31%|███       | 78/250 [02:41<05:48,  2.03s/it]\n",
      "Loss: 0.02004:  31%|███       | 78/250 [02:42<05:48,  2.03s/it]\n",
      "Loss: 0.02071:  31%|███       | 78/250 [02:42<05:48,  2.03s/it]\n",
      "Loss: 0.02155:  31%|███       | 78/250 [02:42<05:48,  2.03s/it]\n",
      "Loss: 0.01817:  31%|███       | 78/250 [02:42<05:48,  2.03s/it]\n",
      "Loss: 0.01817:  32%|███▏      | 79/250 [02:42<05:47,  2.03s/it]\n",
      "Loss: 0.02266:  32%|███▏      | 79/250 [02:42<05:47,  2.03s/it]\n",
      "Loss: 0.02128:  32%|███▏      | 79/250 [02:42<05:47,  2.03s/it]\n",
      "Loss: 0.02463:  32%|███▏      | 79/250 [02:42<05:47,  2.03s/it]\n",
      "Loss: 0.02121:  32%|███▏      | 79/250 [02:42<05:47,  2.03s/it]\n",
      "Loss: 0.02115:  32%|███▏      | 79/250 [02:42<05:47,  2.03s/it]\n",
      "Loss: 0.01918:  32%|███▏      | 79/250 [02:42<05:47,  2.03s/it]\n",
      "Loss: 0.02149:  32%|███▏      | 79/250 [02:43<05:47,  2.03s/it]\n",
      "Loss: 0.02041:  32%|███▏      | 79/250 [02:43<05:47,  2.03s/it]\n",
      "Loss: 0.02257:  32%|███▏      | 79/250 [02:43<05:47,  2.03s/it]\n",
      "Loss: 0.02132:  32%|███▏      | 79/250 [02:43<05:47,  2.03s/it]\n",
      "Loss: 0.02415:  32%|███▏      | 79/250 [02:43<05:47,  2.03s/it]\n",
      "Loss: 0.02180:  32%|███▏      | 79/250 [02:43<05:47,  2.03s/it]\n",
      "Loss: 0.02220:  32%|███▏      | 79/250 [02:43<05:47,  2.03s/it]\n",
      "Loss: 0.02394:  32%|███▏      | 79/250 [02:43<05:47,  2.03s/it]\n",
      "Loss: 0.02140:  32%|███▏      | 79/250 [02:43<05:47,  2.03s/it]\n",
      "Loss: 0.01979:  32%|███▏      | 79/250 [02:44<05:47,  2.03s/it]\n",
      "Loss: 0.02360:  32%|███▏      | 79/250 [02:44<05:47,  2.03s/it]\n",
      "Loss: 0.02226:  32%|███▏      | 79/250 [02:44<05:47,  2.03s/it]\n",
      "Loss: 0.01813:  32%|███▏      | 79/250 [02:44<05:47,  2.03s/it]\n",
      "Loss: 0.01813:  32%|███▏      | 80/250 [02:44<05:45,  2.03s/it]\n",
      "Loss: 0.02196:  32%|███▏      | 80/250 [02:44<05:45,  2.03s/it]\n",
      "Loss: 0.02128:  32%|███▏      | 80/250 [02:44<05:45,  2.03s/it]\n",
      "Loss: 0.02434:  32%|███▏      | 80/250 [02:44<05:45,  2.03s/it]\n",
      "Loss: 0.02116:  32%|███▏      | 80/250 [02:44<05:45,  2.03s/it]\n",
      "Loss: 0.02068:  32%|███▏      | 80/250 [02:44<05:45,  2.03s/it]\n",
      "Loss: 0.01936:  32%|███▏      | 80/250 [02:44<05:45,  2.03s/it]\n",
      "Loss: 0.02157:  32%|███▏      | 80/250 [02:45<05:45,  2.03s/it]\n",
      "Loss: 0.02439:  32%|███▏      | 80/250 [02:45<05:45,  2.03s/it]\n",
      "Loss: 0.02141:  32%|███▏      | 80/250 [02:45<05:45,  2.03s/it]\n",
      "Loss: 0.02370:  32%|███▏      | 80/250 [02:45<05:45,  2.03s/it]\n",
      "Loss: 0.02185:  32%|███▏      | 80/250 [02:45<05:45,  2.03s/it]\n",
      "Loss: 0.02383:  32%|███▏      | 80/250 [02:45<05:45,  2.03s/it]\n",
      "Loss: 0.02071:  32%|███▏      | 80/250 [02:45<05:45,  2.03s/it]\n",
      "Loss: 0.02170:  32%|███▏      | 80/250 [02:45<05:45,  2.03s/it]\n",
      "Loss: 0.01950:  32%|███▏      | 80/250 [02:45<05:45,  2.03s/it]\n",
      "Loss: 0.01938:  32%|███▏      | 80/250 [02:46<05:45,  2.03s/it]\n",
      "Loss: 0.02187:  32%|███▏      | 80/250 [02:46<05:45,  2.03s/it]\n",
      "Loss: 0.01837:  32%|███▏      | 80/250 [02:46<05:45,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 62.64it/s]\u001b[A\n",
      "Loss: 0.01837:  32%|███▏      | 81/250 [02:46<05:41,  2.02s/it]\n",
      "Loss: 0.02220:  32%|███▏      | 81/250 [02:46<05:41,  2.02s/it]\n",
      "Loss: 0.02110:  32%|███▏      | 81/250 [02:46<05:41,  2.02s/it]\n",
      "Loss: 0.02439:  32%|███▏      | 81/250 [02:46<05:41,  2.02s/it]\n",
      "Loss: 0.02117:  32%|███▏      | 81/250 [02:46<05:41,  2.02s/it]\n",
      "Loss: 0.02125:  32%|███▏      | 81/250 [02:46<05:41,  2.02s/it]\n",
      "Loss: 0.01917:  32%|███▏      | 81/250 [02:46<05:41,  2.02s/it]\n",
      "Loss: 0.02127:  32%|███▏      | 81/250 [02:47<05:41,  2.02s/it]\n",
      "Loss: 0.02042:  32%|███▏      | 81/250 [02:47<05:41,  2.02s/it]\n",
      "Loss: 0.02244:  32%|███▏      | 81/250 [02:47<05:41,  2.02s/it]\n",
      "Loss: 0.02186:  32%|███▏      | 81/250 [02:47<05:41,  2.02s/it]\n",
      "Loss: 0.02005:  32%|███▏      | 81/250 [02:47<05:41,  2.02s/it]\n",
      "Loss: 0.02447:  32%|███▏      | 81/250 [02:47<05:41,  2.02s/it]\n",
      "Loss: 0.02411:  32%|███▏      | 81/250 [02:47<05:41,  2.02s/it]\n",
      "Loss: 0.02091:  32%|███▏      | 81/250 [02:47<05:41,  2.02s/it]\n",
      "Loss: 0.02069:  32%|███▏      | 81/250 [02:47<05:41,  2.02s/it]\n",
      "Loss: 0.02158:  32%|███▏      | 81/250 [02:48<05:41,  2.02s/it]\n",
      "Loss: 0.02292:  32%|███▏      | 81/250 [02:48<05:41,  2.02s/it]\n",
      "Loss: 0.02120:  32%|███▏      | 81/250 [02:48<05:41,  2.02s/it]\n",
      "Loss: 0.01812:  32%|███▏      | 81/250 [02:48<05:41,  2.02s/it]\n",
      "Loss: 0.01812:  33%|███▎      | 82/250 [02:48<05:41,  2.03s/it]\n",
      "Loss: 0.02215:  33%|███▎      | 82/250 [02:48<05:41,  2.03s/it]\n",
      "Loss: 0.02092:  33%|███▎      | 82/250 [02:48<05:41,  2.03s/it]\n",
      "Loss: 0.02388:  33%|███▎      | 82/250 [02:48<05:41,  2.03s/it]\n",
      "Loss: 0.02125:  33%|███▎      | 82/250 [02:48<05:41,  2.03s/it]\n",
      "Loss: 0.02080:  33%|███▎      | 82/250 [02:48<05:41,  2.03s/it]\n",
      "Loss: 0.02072:  33%|███▎      | 82/250 [02:49<05:41,  2.03s/it]\n",
      "Loss: 0.02146:  33%|███▎      | 82/250 [02:49<05:41,  2.03s/it]\n",
      "Loss: 0.02243:  33%|███▎      | 82/250 [02:49<05:41,  2.03s/it]\n",
      "Loss: 0.02106:  33%|███▎      | 82/250 [02:49<05:41,  2.03s/it]\n",
      "Loss: 0.02308:  33%|███▎      | 82/250 [02:49<05:41,  2.03s/it]\n",
      "Loss: 0.02242:  33%|███▎      | 82/250 [02:49<05:41,  2.03s/it]\n",
      "Loss: 0.02370:  33%|███▎      | 82/250 [02:49<05:41,  2.03s/it]\n",
      "Loss: 0.02053:  33%|███▎      | 82/250 [02:49<05:41,  2.03s/it]\n",
      "Loss: 0.02137:  33%|███▎      | 82/250 [02:49<05:41,  2.03s/it]\n",
      "Loss: 0.01953:  33%|███▎      | 82/250 [02:50<05:41,  2.03s/it]\n",
      "Loss: 0.01954:  33%|███▎      | 82/250 [02:50<05:41,  2.03s/it]\n",
      "Loss: 0.02181:  33%|███▎      | 82/250 [02:50<05:41,  2.03s/it]\n",
      "Loss: 0.01813:  33%|███▎      | 82/250 [02:50<05:41,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 61.37it/s]\u001b[A\n",
      "Loss: 0.01813:  33%|███▎      | 83/250 [02:50<05:38,  2.03s/it]\n",
      "Loss: 0.02252:  33%|███▎      | 83/250 [02:50<05:38,  2.03s/it]\n",
      "Loss: 0.02088:  33%|███▎      | 83/250 [02:50<05:38,  2.03s/it]\n",
      "Loss: 0.02382:  33%|███▎      | 83/250 [02:50<05:38,  2.03s/it]\n",
      "Loss: 0.02154:  33%|███▎      | 83/250 [02:50<05:38,  2.03s/it]\n",
      "Loss: 0.02051:  33%|███▎      | 83/250 [02:50<05:38,  2.03s/it]\n",
      "Loss: 0.01908:  33%|███▎      | 83/250 [02:51<05:38,  2.03s/it]\n",
      "Loss: 0.02126:  33%|███▎      | 83/250 [02:51<05:38,  2.03s/it]\n",
      "Loss: 0.02411:  33%|███▎      | 83/250 [02:51<05:38,  2.03s/it]\n",
      "Loss: 0.02126:  33%|███▎      | 83/250 [02:51<05:38,  2.03s/it]\n",
      "Loss: 0.02290:  33%|███▎      | 83/250 [02:51<05:38,  2.03s/it]\n",
      "Loss: 0.02186:  33%|███▎      | 83/250 [02:51<05:38,  2.03s/it]\n",
      "Loss: 0.02353:  33%|███▎      | 83/250 [02:51<05:38,  2.03s/it]\n",
      "Loss: 0.02072:  33%|███▎      | 83/250 [02:51<05:38,  2.03s/it]\n",
      "Loss: 0.02147:  33%|███▎      | 83/250 [02:51<05:38,  2.03s/it]\n",
      "Loss: 0.01976:  33%|███▎      | 83/250 [02:52<05:38,  2.03s/it]\n",
      "Loss: 0.01933:  33%|███▎      | 83/250 [02:52<05:38,  2.03s/it]\n",
      "Loss: 0.02187:  33%|███▎      | 83/250 [02:52<05:38,  2.03s/it]\n",
      "Loss: 0.01842:  33%|███▎      | 83/250 [02:52<05:38,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 63.42it/s]\u001b[A\n",
      "Loss: 0.01842:  34%|███▎      | 84/250 [02:52<05:35,  2.02s/it]\n",
      "Loss: 0.02203:  34%|███▎      | 84/250 [02:52<05:35,  2.02s/it]\n",
      "Loss: 0.02078:  34%|███▎      | 84/250 [02:52<05:35,  2.02s/it]\n",
      "Loss: 0.02400:  34%|███▎      | 84/250 [02:52<05:35,  2.02s/it]\n",
      "Loss: 0.02118:  34%|███▎      | 84/250 [02:52<05:35,  2.02s/it]\n",
      "Loss: 0.02048:  34%|███▎      | 84/250 [02:52<05:35,  2.02s/it]\n",
      "Loss: 0.02082:  34%|███▎      | 84/250 [02:53<05:35,  2.02s/it]\n",
      "Loss: 0.02150:  34%|███▎      | 84/250 [02:53<05:35,  2.02s/it]\n",
      "Loss: 0.02156:  34%|███▎      | 84/250 [02:53<05:35,  2.02s/it]\n",
      "Loss: 0.01910:  34%|███▎      | 84/250 [02:53<05:35,  2.02s/it]\n",
      "Loss: 0.02335:  34%|███▎      | 84/250 [02:53<05:35,  2.02s/it]\n",
      "Loss: 0.02400:  34%|███▎      | 84/250 [02:53<05:35,  2.02s/it]\n",
      "Loss: 0.02211:  34%|███▎      | 84/250 [02:53<05:35,  2.02s/it]\n",
      "Loss: 0.02210:  34%|███▎      | 84/250 [02:53<05:35,  2.02s/it]\n",
      "Loss: 0.02022:  34%|███▎      | 84/250 [02:53<05:35,  2.02s/it]\n",
      "Loss: 0.02133:  34%|███▎      | 84/250 [02:54<05:35,  2.02s/it]\n",
      "Loss: 0.02290:  34%|███▎      | 84/250 [02:54<05:35,  2.02s/it]\n",
      "Loss: 0.02149:  34%|███▎      | 84/250 [02:54<05:35,  2.02s/it]\n",
      "Loss: 0.01813:  34%|███▎      | 84/250 [02:54<05:35,  2.02s/it]\n",
      "Loss: 0.01813:  34%|███▍      | 85/250 [02:54<05:32,  2.01s/it]\n",
      "Loss: 0.02199:  34%|███▍      | 85/250 [02:54<05:32,  2.01s/it]\n",
      "Loss: 0.02109:  34%|███▍      | 85/250 [02:54<05:32,  2.01s/it]\n",
      "Loss: 0.02396:  34%|███▍      | 85/250 [02:54<05:32,  2.01s/it]\n",
      "Loss: 0.02121:  34%|███▍      | 85/250 [02:54<05:32,  2.01s/it]\n",
      "Loss: 0.02065:  34%|███▍      | 85/250 [02:54<05:32,  2.01s/it]\n",
      "Loss: 0.02089:  34%|███▍      | 85/250 [02:55<05:32,  2.01s/it]\n",
      "Loss: 0.02158:  34%|███▍      | 85/250 [02:55<05:32,  2.01s/it]\n",
      "Loss: 0.02147:  34%|███▍      | 85/250 [02:55<05:32,  2.01s/it]\n",
      "Loss: 0.02131:  34%|███▍      | 85/250 [02:55<05:32,  2.01s/it]\n",
      "Loss: 0.02288:  34%|███▍      | 85/250 [02:55<05:32,  2.01s/it]\n",
      "Loss: 0.02188:  34%|███▍      | 85/250 [02:55<05:32,  2.01s/it]\n",
      "Loss: 0.02381:  34%|███▍      | 85/250 [02:55<05:32,  2.01s/it]\n",
      "Loss: 0.02044:  34%|███▍      | 85/250 [02:55<05:32,  2.01s/it]\n",
      "Loss: 0.02190:  34%|███▍      | 85/250 [02:55<05:32,  2.01s/it]\n",
      "Loss: 0.02019:  34%|███▍      | 85/250 [02:56<05:32,  2.01s/it]\n",
      "Loss: 0.01909:  34%|███▍      | 85/250 [02:56<05:32,  2.01s/it]\n",
      "Loss: 0.02177:  34%|███▍      | 85/250 [02:56<05:32,  2.01s/it]\n",
      "Loss: 0.01802:  34%|███▍      | 85/250 [02:56<05:32,  2.01s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 63.61it/s]\u001b[A\n",
      "Loss: 0.01802:  34%|███▍      | 86/250 [02:56<05:29,  2.01s/it]\n",
      "Loss: 0.02202:  34%|███▍      | 86/250 [02:56<05:29,  2.01s/it]\n",
      "Loss: 0.02117:  34%|███▍      | 86/250 [02:56<05:29,  2.01s/it]\n",
      "Loss: 0.02410:  34%|███▍      | 86/250 [02:56<05:29,  2.01s/it]\n",
      "Loss: 0.02119:  34%|███▍      | 86/250 [02:56<05:29,  2.01s/it]\n",
      "Loss: 0.02063:  34%|███▍      | 86/250 [02:56<05:29,  2.01s/it]\n",
      "Loss: 0.02080:  34%|███▍      | 86/250 [02:57<05:29,  2.01s/it]\n",
      "Loss: 0.02137:  34%|███▍      | 86/250 [02:57<05:29,  2.01s/it]\n",
      "Loss: 0.02164:  34%|███▍      | 86/250 [02:57<05:29,  2.01s/it]\n",
      "Loss: 0.02117:  34%|███▍      | 86/250 [02:57<05:29,  2.01s/it]\n",
      "Loss: 0.02308:  34%|███▍      | 86/250 [02:57<05:29,  2.01s/it]\n",
      "Loss: 0.02198:  34%|███▍      | 86/250 [02:57<05:29,  2.01s/it]\n",
      "Loss: 0.02355:  34%|███▍      | 86/250 [02:57<05:29,  2.01s/it]\n",
      "Loss: 0.02055:  34%|███▍      | 86/250 [02:57<05:29,  2.01s/it]\n",
      "Loss: 0.02190:  34%|███▍      | 86/250 [02:57<05:29,  2.01s/it]\n",
      "Loss: 0.01940:  34%|███▍      | 86/250 [02:58<05:29,  2.01s/it]\n",
      "Loss: 0.01923:  34%|███▍      | 86/250 [02:58<05:29,  2.01s/it]\n",
      "Loss: 0.02157:  34%|███▍      | 86/250 [02:58<05:29,  2.01s/it]\n",
      "Loss: 0.01828:  34%|███▍      | 86/250 [02:58<05:29,  2.01s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 62.23it/s]\u001b[A\n",
      "Loss: 0.01828:  35%|███▍      | 87/250 [02:58<05:27,  2.01s/it]\n",
      "Loss: 0.02184:  35%|███▍      | 87/250 [02:58<05:27,  2.01s/it]\n",
      "Loss: 0.02124:  35%|███▍      | 87/250 [02:58<05:27,  2.01s/it]\n",
      "Loss: 0.02393:  35%|███▍      | 87/250 [02:58<05:27,  2.01s/it]\n",
      "Loss: 0.02122:  35%|███▍      | 87/250 [02:58<05:27,  2.01s/it]\n",
      "Loss: 0.02160:  35%|███▍      | 87/250 [02:58<05:27,  2.01s/it]\n",
      "Loss: 0.02213:  35%|███▍      | 87/250 [02:59<05:27,  2.01s/it]\n",
      "Loss: 0.02025:  35%|███▍      | 87/250 [02:59<05:27,  2.01s/it]\n",
      "Loss: 0.02033:  35%|███▍      | 87/250 [02:59<05:27,  2.01s/it]\n",
      "Loss: 0.02211:  35%|███▍      | 87/250 [02:59<05:27,  2.01s/it]\n",
      "Loss: 0.02130:  35%|███▍      | 87/250 [02:59<05:27,  2.01s/it]\n",
      "Loss: 0.02425:  35%|███▍      | 87/250 [02:59<05:27,  2.01s/it]\n",
      "Loss: 0.02185:  35%|███▍      | 87/250 [02:59<05:27,  2.01s/it]\n",
      "Loss: 0.02227:  35%|███▍      | 87/250 [02:59<05:27,  2.01s/it]\n",
      "Loss: 0.02361:  35%|███▍      | 87/250 [02:59<05:27,  2.01s/it]\n",
      "Loss: 0.02146:  35%|███▍      | 87/250 [03:00<05:27,  2.01s/it]\n",
      "Loss: 0.02102:  35%|███▍      | 87/250 [03:00<05:27,  2.01s/it]\n",
      "Loss: 0.02294:  35%|███▍      | 87/250 [03:00<05:27,  2.01s/it]\n",
      "Loss: 0.02229:  35%|███▍      | 87/250 [03:00<05:27,  2.01s/it]\n",
      "Loss: 0.01811:  35%|███▍      | 87/250 [03:00<05:27,  2.01s/it]\n",
      "Loss: 0.01811:  35%|███▌      | 88/250 [03:00<05:25,  2.01s/it]\n",
      "Loss: 0.02185:  35%|███▌      | 88/250 [03:00<05:25,  2.01s/it]\n",
      "Loss: 0.02161:  35%|███▌      | 88/250 [03:00<05:25,  2.01s/it]\n",
      "Loss: 0.02385:  35%|███▌      | 88/250 [03:00<05:25,  2.01s/it]\n",
      "Loss: 0.02127:  35%|███▌      | 88/250 [03:00<05:25,  2.01s/it]\n",
      "Loss: 0.02052:  35%|███▌      | 88/250 [03:00<05:25,  2.01s/it]\n",
      "Loss: 0.02123:  35%|███▌      | 88/250 [03:01<05:25,  2.01s/it]\n",
      "Loss: 0.02156:  35%|███▌      | 88/250 [03:01<05:25,  2.01s/it]\n",
      "Loss: 0.02163:  35%|███▌      | 88/250 [03:01<05:25,  2.01s/it]\n",
      "Loss: 0.01911:  35%|███▌      | 88/250 [03:01<05:25,  2.01s/it]\n",
      "Loss: 0.02331:  35%|███▌      | 88/250 [03:01<05:25,  2.01s/it]\n",
      "Loss: 0.02442:  35%|███▌      | 88/250 [03:01<05:25,  2.01s/it]\n",
      "Loss: 0.02235:  35%|███▌      | 88/250 [03:01<05:25,  2.01s/it]\n",
      "Loss: 0.02190:  35%|███▌      | 88/250 [03:01<05:25,  2.01s/it]\n",
      "Loss: 0.02026:  35%|███▌      | 88/250 [03:01<05:25,  2.01s/it]\n",
      "Loss: 0.02128:  35%|███▌      | 88/250 [03:02<05:25,  2.01s/it]\n",
      "Loss: 0.02292:  35%|███▌      | 88/250 [03:02<05:25,  2.01s/it]\n",
      "Loss: 0.02145:  35%|███▌      | 88/250 [03:02<05:25,  2.01s/it]\n",
      "Loss: 0.01813:  35%|███▌      | 88/250 [03:02<05:25,  2.01s/it]\n",
      "Loss: 0.01813:  36%|███▌      | 89/250 [03:02<05:24,  2.02s/it]\n",
      "Loss: 0.02199:  36%|███▌      | 89/250 [03:02<05:24,  2.02s/it]\n",
      "Loss: 0.02099:  36%|███▌      | 89/250 [03:02<05:24,  2.02s/it]\n",
      "Loss: 0.02367:  36%|███▌      | 89/250 [03:02<05:24,  2.02s/it]\n",
      "Loss: 0.02129:  36%|███▌      | 89/250 [03:02<05:24,  2.02s/it]\n",
      "Loss: 0.02045:  36%|███▌      | 89/250 [03:03<05:24,  2.02s/it]\n",
      "Loss: 0.02076:  36%|███▌      | 89/250 [03:03<05:24,  2.02s/it]\n",
      "Loss: 0.02145:  36%|███▌      | 89/250 [03:03<05:24,  2.02s/it]\n",
      "Loss: 0.02169:  36%|███▌      | 89/250 [03:03<05:24,  2.02s/it]\n",
      "Loss: 0.01914:  36%|███▌      | 89/250 [03:03<05:24,  2.02s/it]\n",
      "Loss: 0.02419:  36%|███▌      | 89/250 [03:03<05:24,  2.02s/it]\n",
      "Loss: 0.02408:  36%|███▌      | 89/250 [03:03<05:24,  2.02s/it]\n",
      "Loss: 0.02195:  36%|███▌      | 89/250 [03:03<05:24,  2.02s/it]\n",
      "Loss: 0.02163:  36%|███▌      | 89/250 [03:03<05:24,  2.02s/it]\n",
      "Loss: 0.02038:  36%|███▌      | 89/250 [03:04<05:24,  2.02s/it]\n",
      "Loss: 0.02136:  36%|███▌      | 89/250 [03:04<05:24,  2.02s/it]\n",
      "Loss: 0.02261:  36%|███▌      | 89/250 [03:04<05:24,  2.02s/it]\n",
      "Loss: 0.02115:  36%|███▌      | 89/250 [03:04<05:24,  2.02s/it]\n",
      "Loss: 0.01801:  36%|███▌      | 89/250 [03:04<05:24,  2.02s/it]\n",
      "Loss: 0.01801:  36%|███▌      | 90/250 [03:04<05:23,  2.02s/it]\n",
      "Loss: 0.02180:  36%|███▌      | 90/250 [03:04<05:23,  2.02s/it]\n",
      "Loss: 0.01920:  36%|███▌      | 90/250 [03:04<05:23,  2.02s/it]\n",
      "Loss: 0.02332:  36%|███▌      | 90/250 [03:04<05:23,  2.02s/it]\n",
      "Loss: 0.01982:  36%|███▌      | 90/250 [03:04<05:23,  2.02s/it]\n",
      "Loss: 0.01819:  36%|███▌      | 90/250 [03:05<05:23,  2.02s/it]\n",
      "Loss: 0.02207:  36%|███▌      | 90/250 [03:05<05:23,  2.02s/it]\n",
      "Loss: 0.02006:  36%|███▌      | 90/250 [03:05<05:23,  2.02s/it]\n",
      "Loss: 0.02038:  36%|███▌      | 90/250 [03:05<05:23,  2.02s/it]\n",
      "Loss: 0.02216:  36%|███▌      | 90/250 [03:05<05:23,  2.02s/it]\n",
      "Loss: 0.02111:  36%|███▌      | 90/250 [03:05<05:23,  2.02s/it]\n",
      "Loss: 0.02411:  36%|███▌      | 90/250 [03:05<05:23,  2.02s/it]\n",
      "Loss: 0.02139:  36%|███▌      | 90/250 [03:05<05:23,  2.02s/it]\n",
      "Loss: 0.02211:  36%|███▌      | 90/250 [03:05<05:23,  2.02s/it]\n",
      "Loss: 0.02380:  36%|███▌      | 90/250 [03:06<05:23,  2.02s/it]\n",
      "Loss: 0.02117:  36%|███▌      | 90/250 [03:06<05:23,  2.02s/it]\n",
      "Loss: 0.01970:  36%|███▌      | 90/250 [03:06<05:23,  2.02s/it]\n",
      "Loss: 0.02285:  36%|███▌      | 90/250 [03:06<05:23,  2.02s/it]\n",
      "Loss: 0.02219:  36%|███▌      | 90/250 [03:06<05:23,  2.02s/it]\n",
      "Loss: 0.01848:  36%|███▌      | 90/250 [03:06<05:23,  2.02s/it]\n",
      "Loss: 0.01848:  36%|███▋      | 91/250 [03:06<05:20,  2.01s/it]\n",
      "Loss: 0.02178:  36%|███▋      | 91/250 [03:06<05:20,  2.01s/it]\n",
      "Loss: 0.02094:  36%|███▋      | 91/250 [03:06<05:20,  2.01s/it]\n",
      "Loss: 0.02365:  36%|███▋      | 91/250 [03:06<05:20,  2.01s/it]\n",
      "Loss: 0.02123:  36%|███▋      | 91/250 [03:06<05:20,  2.01s/it]\n",
      "Loss: 0.02078:  36%|███▋      | 91/250 [03:07<05:20,  2.01s/it]\n",
      "Loss: 0.01888:  36%|███▋      | 91/250 [03:07<05:20,  2.01s/it]\n",
      "Loss: 0.02012:  36%|███▋      | 91/250 [03:07<05:20,  2.01s/it]\n",
      "Loss: 0.02015:  36%|███▋      | 91/250 [03:07<05:20,  2.01s/it]\n",
      "Loss: 0.02193:  36%|███▋      | 91/250 [03:07<05:20,  2.01s/it]\n",
      "Loss: 0.02177:  36%|███▋      | 91/250 [03:07<05:20,  2.01s/it]\n",
      "Loss: 0.02007:  36%|███▋      | 91/250 [03:07<05:20,  2.01s/it]\n",
      "Loss: 0.02449:  36%|███▋      | 91/250 [03:07<05:20,  2.01s/it]\n",
      "Loss: 0.02085:  36%|███▋      | 91/250 [03:07<05:20,  2.01s/it]\n",
      "Loss: 0.02371:  36%|███▋      | 91/250 [03:07<05:20,  2.01s/it]\n",
      "Loss: 0.01955:  36%|███▋      | 91/250 [03:08<05:20,  2.01s/it]\n",
      "Loss: 0.02017:  36%|███▋      | 91/250 [03:08<05:20,  2.01s/it]\n",
      "Loss: 0.02007:  36%|███▋      | 91/250 [03:08<05:20,  2.01s/it]\n",
      "Loss: 0.02112:  36%|███▋      | 91/250 [03:08<05:20,  2.01s/it]\n",
      "Loss: 0.01848:  36%|███▋      | 91/250 [03:08<05:20,  2.01s/it]\n",
      "Loss: 0.01848:  37%|███▋      | 92/250 [03:08<05:18,  2.01s/it]\n",
      "Loss: 0.02248:  37%|███▋      | 92/250 [03:08<05:18,  2.01s/it]\n",
      "Loss: 0.01907:  37%|███▋      | 92/250 [03:08<05:18,  2.01s/it]\n",
      "Loss: 0.01908:  37%|███▋      | 92/250 [03:08<05:18,  2.01s/it]\n",
      "Loss: 0.02334:  37%|███▋      | 92/250 [03:08<05:18,  2.01s/it]\n",
      "Loss: 0.01795:  37%|███▋      | 92/250 [03:09<05:18,  2.01s/it]\n",
      "Loss: 0.02193:  37%|███▋      | 92/250 [03:09<05:18,  2.01s/it]\n",
      "Loss: 0.02330:  37%|███▋      | 92/250 [03:09<05:18,  2.01s/it]\n",
      "Loss: 0.02167:  37%|███▋      | 92/250 [03:09<05:18,  2.01s/it]\n",
      "Loss: 0.02241:  37%|███▋      | 92/250 [03:09<05:18,  2.01s/it]\n",
      "Loss: 0.02204:  37%|███▋      | 92/250 [03:09<05:18,  2.01s/it]\n",
      "Loss: 0.02023:  37%|███▋      | 92/250 [03:09<05:18,  2.01s/it]\n",
      "Loss: 0.02113:  37%|███▋      | 92/250 [03:09<05:18,  2.01s/it]\n",
      "Loss: 0.02164:  37%|███▋      | 92/250 [03:09<05:18,  2.01s/it]\n",
      "Loss: 0.01915:  37%|███▋      | 92/250 [03:09<05:18,  2.01s/it]\n",
      "Loss: 0.02311:  37%|███▋      | 92/250 [03:10<05:18,  2.01s/it]\n",
      "Loss: 0.02468:  37%|███▋      | 92/250 [03:10<05:18,  2.01s/it]\n",
      "Loss: 0.02413:  37%|███▋      | 92/250 [03:10<05:18,  2.01s/it]\n",
      "Loss: 0.02020:  37%|███▋      | 92/250 [03:10<05:18,  2.01s/it]\n",
      "Loss: 0.01822:  37%|███▋      | 92/250 [03:10<05:18,  2.01s/it]\n",
      "Loss: 0.01822:  37%|███▋      | 93/250 [03:10<05:15,  2.01s/it]\n",
      "Loss: 0.02168:  37%|███▋      | 93/250 [03:10<05:15,  2.01s/it]\n",
      "Loss: 0.02077:  37%|███▋      | 93/250 [03:10<05:15,  2.01s/it]\n",
      "Loss: 0.02484:  37%|███▋      | 93/250 [03:10<05:15,  2.01s/it]\n",
      "Loss: 0.01953:  37%|███▋      | 93/250 [03:10<05:15,  2.01s/it]\n",
      "Loss: 0.02119:  37%|███▋      | 93/250 [03:11<05:15,  2.01s/it]\n",
      "Loss: 0.01906:  37%|███▋      | 93/250 [03:11<05:15,  2.01s/it]\n",
      "Loss: 0.02014:  37%|███▋      | 93/250 [03:11<05:15,  2.01s/it]\n",
      "Loss: 0.02018:  37%|███▋      | 93/250 [03:11<05:15,  2.01s/it]\n",
      "Loss: 0.02206:  37%|███▋      | 93/250 [03:11<05:15,  2.01s/it]\n",
      "Loss: 0.02108:  37%|███▋      | 93/250 [03:11<05:15,  2.01s/it]\n",
      "Loss: 0.02408:  37%|███▋      | 93/250 [03:11<05:15,  2.01s/it]\n",
      "Loss: 0.02184:  37%|███▋      | 93/250 [03:11<05:15,  2.01s/it]\n",
      "Loss: 0.02185:  37%|███▋      | 93/250 [03:11<05:15,  2.01s/it]\n",
      "Loss: 0.02382:  37%|███▋      | 93/250 [03:12<05:15,  2.01s/it]\n",
      "Loss: 0.02133:  37%|███▋      | 93/250 [03:12<05:15,  2.01s/it]\n",
      "Loss: 0.01979:  37%|███▋      | 93/250 [03:12<05:15,  2.01s/it]\n",
      "Loss: 0.02320:  37%|███▋      | 93/250 [03:12<05:15,  2.01s/it]\n",
      "Loss: 0.02240:  37%|███▋      | 93/250 [03:12<05:15,  2.01s/it]\n",
      "Loss: 0.01807:  37%|███▋      | 93/250 [03:12<05:15,  2.01s/it]\n",
      "Loss: 0.01807:  38%|███▊      | 94/250 [03:12<05:13,  2.01s/it]\n",
      "Loss: 0.02187:  38%|███▊      | 94/250 [03:12<05:13,  2.01s/it]\n",
      "Loss: 0.02090:  38%|███▊      | 94/250 [03:12<05:13,  2.01s/it]\n",
      "Loss: 0.02360:  38%|███▊      | 94/250 [03:12<05:13,  2.01s/it]\n",
      "Loss: 0.02115:  38%|███▊      | 94/250 [03:12<05:13,  2.01s/it]\n",
      "Loss: 0.02046:  38%|███▊      | 94/250 [03:13<05:13,  2.01s/it]\n",
      "Loss: 0.01929:  38%|███▊      | 94/250 [03:13<05:13,  2.01s/it]\n",
      "Loss: 0.02116:  38%|███▊      | 94/250 [03:13<05:13,  2.01s/it]\n",
      "Loss: 0.02385:  38%|███▊      | 94/250 [03:13<05:13,  2.01s/it]\n",
      "Loss: 0.02115:  38%|███▊      | 94/250 [03:13<05:13,  2.01s/it]\n",
      "Loss: 0.02284:  38%|███▊      | 94/250 [03:13<05:13,  2.01s/it]\n",
      "Loss: 0.02215:  38%|███▊      | 94/250 [03:13<05:13,  2.01s/it]\n",
      "Loss: 0.02354:  38%|███▊      | 94/250 [03:13<05:13,  2.01s/it]\n",
      "Loss: 0.02045:  38%|███▊      | 94/250 [03:13<05:13,  2.01s/it]\n",
      "Loss: 0.02102:  38%|███▊      | 94/250 [03:14<05:13,  2.01s/it]\n",
      "Loss: 0.01967:  38%|███▊      | 94/250 [03:14<05:13,  2.01s/it]\n",
      "Loss: 0.01928:  38%|███▊      | 94/250 [03:14<05:13,  2.01s/it]\n",
      "Loss: 0.02163:  38%|███▊      | 94/250 [03:14<05:13,  2.01s/it]\n",
      "Loss: 0.01826:  38%|███▊      | 94/250 [03:14<05:13,  2.01s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 62.62it/s]\u001b[A\n",
      "Loss: 0.01826:  38%|███▊      | 95/250 [03:14<05:11,  2.01s/it]\n",
      "Loss: 0.02063:  38%|███▊      | 95/250 [03:14<05:11,  2.01s/it]\n",
      "Loss: 0.01979:  38%|███▊      | 95/250 [03:14<05:11,  2.01s/it]\n",
      "Loss: 0.01931:  38%|███▊      | 95/250 [03:14<05:11,  2.01s/it]\n",
      "Loss: 0.02301:  38%|███▊      | 95/250 [03:14<05:11,  2.01s/it]\n",
      "Loss: 0.01817:  38%|███▊      | 95/250 [03:15<05:11,  2.01s/it]\n",
      "Loss: 0.02192:  38%|███▊      | 95/250 [03:15<05:11,  2.01s/it]\n",
      "Loss: 0.02025:  38%|███▊      | 95/250 [03:15<05:11,  2.01s/it]\n",
      "Loss: 0.02022:  38%|███▊      | 95/250 [03:15<05:11,  2.01s/it]\n",
      "Loss: 0.02250:  38%|███▊      | 95/250 [03:15<05:11,  2.01s/it]\n",
      "Loss: 0.02099:  38%|███▊      | 95/250 [03:15<05:11,  2.01s/it]\n",
      "Loss: 0.02420:  38%|███▊      | 95/250 [03:15<05:11,  2.01s/it]\n",
      "Loss: 0.02163:  38%|███▊      | 95/250 [03:15<05:11,  2.01s/it]\n",
      "Loss: 0.02191:  38%|███▊      | 95/250 [03:15<05:11,  2.01s/it]\n",
      "Loss: 0.02380:  38%|███▊      | 95/250 [03:16<05:11,  2.01s/it]\n",
      "Loss: 0.02156:  38%|███▊      | 95/250 [03:16<05:11,  2.01s/it]\n",
      "Loss: 0.02007:  38%|███▊      | 95/250 [03:16<05:11,  2.01s/it]\n",
      "Loss: 0.02313:  38%|███▊      | 95/250 [03:16<05:11,  2.01s/it]\n",
      "Loss: 0.02267:  38%|███▊      | 95/250 [03:16<05:11,  2.01s/it]\n",
      "Loss: 0.01803:  38%|███▊      | 95/250 [03:16<05:11,  2.01s/it]\n",
      "Loss: 0.01803:  38%|███▊      | 96/250 [03:16<05:09,  2.01s/it]\n",
      "Loss: 0.02054:  38%|███▊      | 96/250 [03:16<05:09,  2.01s/it]\n",
      "Loss: 0.01898:  38%|███▊      | 96/250 [03:16<05:09,  2.01s/it]\n",
      "Loss: 0.02324:  38%|███▊      | 96/250 [03:16<05:09,  2.01s/it]\n",
      "Loss: 0.02346:  38%|███▊      | 96/250 [03:16<05:09,  2.01s/it]\n",
      "Loss: 0.01801:  38%|███▊      | 96/250 [03:17<05:09,  2.01s/it]\n",
      "Loss: 0.02197:  38%|███▊      | 96/250 [03:17<05:09,  2.01s/it]\n",
      "Loss: 0.02310:  38%|███▊      | 96/250 [03:17<05:09,  2.01s/it]\n",
      "Loss: 0.02147:  38%|███▊      | 96/250 [03:17<05:09,  2.01s/it]\n",
      "Loss: 0.02191:  38%|███▊      | 96/250 [03:17<05:09,  2.01s/it]\n",
      "Loss: 0.02163:  38%|███▊      | 96/250 [03:17<05:09,  2.01s/it]\n",
      "Loss: 0.02092:  38%|███▊      | 96/250 [03:17<05:09,  2.01s/it]\n",
      "Loss: 0.01965:  38%|███▊      | 96/250 [03:17<05:09,  2.01s/it]\n",
      "Loss: 0.02421:  38%|███▊      | 96/250 [03:17<05:09,  2.01s/it]\n",
      "Loss: 0.02172:  38%|███▊      | 96/250 [03:18<05:09,  2.01s/it]\n",
      "Loss: 0.02045:  38%|███▊      | 96/250 [03:18<05:09,  2.01s/it]\n",
      "Loss: 0.02433:  38%|███▊      | 96/250 [03:18<05:09,  2.01s/it]\n",
      "Loss: 0.01957:  38%|███▊      | 96/250 [03:18<05:09,  2.01s/it]\n",
      "Loss: 0.02099:  38%|███▊      | 96/250 [03:18<05:09,  2.01s/it]\n",
      "Loss: 0.01783:  38%|███▊      | 96/250 [03:18<05:09,  2.01s/it]\n",
      "Loss: 0.01783:  39%|███▉      | 97/250 [03:18<05:09,  2.02s/it]\n",
      "Loss: 0.02157:  39%|███▉      | 97/250 [03:18<05:09,  2.02s/it]\n",
      "Loss: 0.02083:  39%|███▉      | 97/250 [03:18<05:09,  2.02s/it]\n",
      "Loss: 0.02413:  39%|███▉      | 97/250 [03:18<05:09,  2.02s/it]\n",
      "Loss: 0.02092:  39%|███▉      | 97/250 [03:19<05:09,  2.02s/it]\n",
      "Loss: 0.02037:  39%|███▉      | 97/250 [03:19<05:09,  2.02s/it]\n",
      "Loss: 0.02043:  39%|███▉      | 97/250 [03:19<05:09,  2.02s/it]\n",
      "Loss: 0.02170:  39%|███▉      | 97/250 [03:19<05:09,  2.02s/it]\n",
      "Loss: 0.02154:  39%|███▉      | 97/250 [03:19<05:09,  2.02s/it]\n",
      "Loss: 0.01889:  39%|███▉      | 97/250 [03:19<05:09,  2.02s/it]\n",
      "Loss: 0.02316:  39%|███▉      | 97/250 [03:19<05:09,  2.02s/it]\n",
      "Loss: 0.02361:  39%|███▉      | 97/250 [03:19<05:09,  2.02s/it]\n",
      "Loss: 0.02171:  39%|███▉      | 97/250 [03:19<05:09,  2.02s/it]\n",
      "Loss: 0.02158:  39%|███▉      | 97/250 [03:20<05:09,  2.02s/it]\n",
      "Loss: 0.02020:  39%|███▉      | 97/250 [03:20<05:09,  2.02s/it]\n",
      "Loss: 0.02124:  39%|███▉      | 97/250 [03:20<05:09,  2.02s/it]\n",
      "Loss: 0.02228:  39%|███▉      | 97/250 [03:20<05:09,  2.02s/it]\n",
      "Loss: 0.02080:  39%|███▉      | 97/250 [03:20<05:09,  2.02s/it]\n",
      "Loss: 0.01813:  39%|███▉      | 97/250 [03:20<05:09,  2.02s/it]\n",
      "Loss: 0.01813:  39%|███▉      | 98/250 [03:20<05:06,  2.02s/it]\n",
      "Loss: 0.02155:  39%|███▉      | 98/250 [03:20<05:06,  2.02s/it]\n",
      "Loss: 0.02092:  39%|███▉      | 98/250 [03:20<05:06,  2.02s/it]\n",
      "Loss: 0.02303:  39%|███▉      | 98/250 [03:20<05:06,  2.02s/it]\n",
      "Loss: 0.01943:  39%|███▉      | 98/250 [03:21<05:06,  2.02s/it]\n",
      "Loss: 0.02088:  39%|███▉      | 98/250 [03:21<05:06,  2.02s/it]\n",
      "Loss: 0.01880:  39%|███▉      | 98/250 [03:21<05:06,  2.02s/it]\n",
      "Loss: 0.02052:  39%|███▉      | 98/250 [03:21<05:06,  2.02s/it]\n",
      "Loss: 0.02324:  39%|███▉      | 98/250 [03:21<05:06,  2.02s/it]\n",
      "Loss: 0.02102:  39%|███▉      | 98/250 [03:21<05:06,  2.02s/it]\n",
      "Loss: 0.02213:  39%|███▉      | 98/250 [03:21<05:06,  2.02s/it]\n",
      "Loss: 0.02178:  39%|███▉      | 98/250 [03:21<05:06,  2.02s/it]\n",
      "Loss: 0.02280:  39%|███▉      | 98/250 [03:21<05:06,  2.02s/it]\n",
      "Loss: 0.01990:  39%|███▉      | 98/250 [03:22<05:06,  2.02s/it]\n",
      "Loss: 0.02076:  39%|███▉      | 98/250 [03:22<05:06,  2.02s/it]\n",
      "Loss: 0.01888:  39%|███▉      | 98/250 [03:22<05:06,  2.02s/it]\n",
      "Loss: 0.01926:  39%|███▉      | 98/250 [03:22<05:06,  2.02s/it]\n",
      "Loss: 0.02137:  39%|███▉      | 98/250 [03:22<05:06,  2.02s/it]\n",
      "Loss: 0.01780:  39%|███▉      | 98/250 [03:22<05:06,  2.02s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 62.14it/s]\u001b[A\n",
      "Loss: 0.01780:  40%|███▉      | 99/250 [03:22<05:03,  2.01s/it]\n",
      "Loss: 0.02096:  40%|███▉      | 99/250 [03:22<05:03,  2.01s/it]\n",
      "Loss: 0.02047:  40%|███▉      | 99/250 [03:22<05:03,  2.01s/it]\n",
      "Loss: 0.02295:  40%|███▉      | 99/250 [03:22<05:03,  2.01s/it]\n",
      "Loss: 0.01971:  40%|███▉      | 99/250 [03:23<05:03,  2.01s/it]\n",
      "Loss: 0.02080:  40%|███▉      | 99/250 [03:23<05:03,  2.01s/it]\n",
      "Loss: 0.01868:  40%|███▉      | 99/250 [03:23<05:03,  2.01s/it]\n",
      "Loss: 0.02008:  40%|███▉      | 99/250 [03:23<05:03,  2.01s/it]\n",
      "Loss: 0.02330:  40%|███▉      | 99/250 [03:23<05:03,  2.01s/it]\n",
      "Loss: 0.02071:  40%|███▉      | 99/250 [03:23<05:03,  2.01s/it]\n",
      "Loss: 0.02301:  40%|███▉      | 99/250 [03:23<05:03,  2.01s/it]\n",
      "Loss: 0.02146:  40%|███▉      | 99/250 [03:23<05:03,  2.01s/it]\n",
      "Loss: 0.02259:  40%|███▉      | 99/250 [03:23<05:03,  2.01s/it]\n",
      "Loss: 0.01950:  40%|███▉      | 99/250 [03:24<05:03,  2.01s/it]\n",
      "Loss: 0.02068:  40%|███▉      | 99/250 [03:24<05:03,  2.01s/it]\n",
      "Loss: 0.01903:  40%|███▉      | 99/250 [03:24<05:03,  2.01s/it]\n",
      "Loss: 0.01872:  40%|███▉      | 99/250 [03:24<05:03,  2.01s/it]\n",
      "Loss: 0.02131:  40%|███▉      | 99/250 [03:24<05:03,  2.01s/it]\n",
      "Loss: 0.01747:  40%|███▉      | 99/250 [03:24<05:03,  2.01s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 61.89it/s]\u001b[A\n",
      "Loss: 0.01747:  40%|████      | 100/250 [03:24<05:01,  2.01s/it]\n",
      "Loss: 0.02096:  40%|████      | 100/250 [03:24<05:01,  2.01s/it]\n",
      "Loss: 0.02037:  40%|████      | 100/250 [03:24<05:01,  2.01s/it]\n",
      "Loss: 0.02278:  40%|████      | 100/250 [03:24<05:01,  2.01s/it]\n",
      "Loss: 0.02087:  40%|████      | 100/250 [03:25<05:01,  2.01s/it]\n",
      "Loss: 0.01987:  40%|████      | 100/250 [03:25<05:01,  2.01s/it]\n",
      "Loss: 0.02087:  40%|████      | 100/250 [03:25<05:01,  2.01s/it]\n",
      "Loss: 0.02062:  40%|████      | 100/250 [03:25<05:01,  2.01s/it]\n",
      "Loss: 0.02114:  40%|████      | 100/250 [03:25<05:01,  2.01s/it]\n",
      "Loss: 0.02110:  40%|████      | 100/250 [03:25<05:01,  2.01s/it]\n",
      "Loss: 0.02163:  40%|████      | 100/250 [03:25<05:01,  2.01s/it]\n",
      "Loss: 0.02146:  40%|████      | 100/250 [03:25<05:01,  2.01s/it]\n",
      "Loss: 0.02244:  40%|████      | 100/250 [03:25<05:01,  2.01s/it]\n",
      "Loss: 0.02015:  40%|████      | 100/250 [03:26<05:01,  2.01s/it]\n",
      "Loss: 0.02026:  40%|████      | 100/250 [03:26<05:01,  2.01s/it]\n",
      "Loss: 0.01897:  40%|████      | 100/250 [03:26<05:01,  2.01s/it]\n",
      "Loss: 0.01894:  40%|████      | 100/250 [03:26<05:01,  2.01s/it]\n",
      "Loss: 0.02074:  40%|████      | 100/250 [03:26<05:01,  2.01s/it]\n",
      "Loss: 0.01747:  40%|████      | 100/250 [03:26<05:01,  2.01s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.12it/s]\u001b[A\n",
      "Loss: 0.01747:  40%|████      | 101/250 [03:26<05:00,  2.02s/it]\n",
      "Loss: 0.02033:  40%|████      | 101/250 [03:26<05:00,  2.02s/it]\n",
      "Loss: 0.01890:  40%|████      | 101/250 [03:26<05:00,  2.02s/it]\n",
      "Loss: 0.01818:  40%|████      | 101/250 [03:26<05:00,  2.02s/it]\n",
      "Loss: 0.02031:  40%|████      | 101/250 [03:27<05:00,  2.02s/it]\n",
      "Loss: 0.02170:  40%|████      | 101/250 [03:27<05:00,  2.02s/it]\n",
      "Loss: 0.02143:  40%|████      | 101/250 [03:27<05:00,  2.02s/it]\n",
      "Loss: 0.02246:  40%|████      | 101/250 [03:27<05:00,  2.02s/it]\n",
      "Loss: 0.02169:  40%|████      | 101/250 [03:27<05:00,  2.02s/it]\n",
      "Loss: 0.02236:  40%|████      | 101/250 [03:27<05:00,  2.02s/it]\n",
      "Loss: 0.02046:  40%|████      | 101/250 [03:27<05:00,  2.02s/it]\n",
      "Loss: 0.02002:  40%|████      | 101/250 [03:27<05:00,  2.02s/it]\n",
      "Loss: 0.02340:  40%|████      | 101/250 [03:27<05:00,  2.02s/it]\n",
      "Loss: 0.02061:  40%|████      | 101/250 [03:28<05:00,  2.02s/it]\n",
      "Loss: 0.02265:  40%|████      | 101/250 [03:28<05:00,  2.02s/it]\n",
      "Loss: 0.01909:  40%|████      | 101/250 [03:28<05:00,  2.02s/it]\n",
      "Loss: 0.01969:  40%|████      | 101/250 [03:28<05:00,  2.02s/it]\n",
      "Loss: 0.01872:  40%|████      | 101/250 [03:28<05:00,  2.02s/it]\n",
      "Loss: 0.02078:  40%|████      | 101/250 [03:28<05:00,  2.02s/it]\n",
      "Loss: 0.01774:  40%|████      | 101/250 [03:28<05:00,  2.02s/it]\n",
      "Loss: 0.01774:  41%|████      | 102/250 [03:28<04:59,  2.02s/it]\n",
      "Loss: 0.02024:  41%|████      | 102/250 [03:28<04:59,  2.02s/it]\n",
      "Loss: 0.01997:  41%|████      | 102/250 [03:28<04:59,  2.02s/it]\n",
      "Loss: 0.02270:  41%|████      | 102/250 [03:28<04:59,  2.02s/it]\n",
      "Loss: 0.02247:  41%|████      | 102/250 [03:29<04:59,  2.02s/it]\n",
      "Loss: 0.01810:  41%|████      | 102/250 [03:29<04:59,  2.02s/it]\n",
      "Loss: 0.02117:  41%|████      | 102/250 [03:29<04:59,  2.02s/it]\n",
      "Loss: 0.01962:  41%|████      | 102/250 [03:29<04:59,  2.02s/it]\n",
      "Loss: 0.01986:  41%|████      | 102/250 [03:29<04:59,  2.02s/it]\n",
      "Loss: 0.02191:  41%|████      | 102/250 [03:29<04:59,  2.02s/it]\n",
      "Loss: 0.02087:  41%|████      | 102/250 [03:29<04:59,  2.02s/it]\n",
      "Loss: 0.02409:  41%|████      | 102/250 [03:29<04:59,  2.02s/it]\n",
      "Loss: 0.02117:  41%|████      | 102/250 [03:29<04:59,  2.02s/it]\n",
      "Loss: 0.02077:  41%|████      | 102/250 [03:30<04:59,  2.02s/it]\n",
      "Loss: 0.02310:  41%|████      | 102/250 [03:30<04:59,  2.02s/it]\n",
      "Loss: 0.02078:  41%|████      | 102/250 [03:30<04:59,  2.02s/it]\n",
      "Loss: 0.01852:  41%|████      | 102/250 [03:30<04:59,  2.02s/it]\n",
      "Loss: 0.02279:  41%|████      | 102/250 [03:30<04:59,  2.02s/it]\n",
      "Loss: 0.02163:  41%|████      | 102/250 [03:30<04:59,  2.02s/it]\n",
      "Loss: 0.01751:  41%|████      | 102/250 [03:30<04:59,  2.02s/it]\n",
      "Loss: 0.01751:  41%|████      | 103/250 [03:30<04:57,  2.02s/it]\n",
      "Loss: 0.02024:  41%|████      | 103/250 [03:30<04:57,  2.02s/it]\n",
      "Loss: 0.01903:  41%|████      | 103/250 [03:30<04:57,  2.02s/it]\n",
      "Loss: 0.02243:  41%|████      | 103/250 [03:31<04:57,  2.02s/it]\n",
      "Loss: 0.01879:  41%|████      | 103/250 [03:31<04:57,  2.02s/it]\n",
      "Loss: 0.02081:  41%|████      | 103/250 [03:31<04:57,  2.02s/it]\n",
      "Loss: 0.01808:  41%|████      | 103/250 [03:31<04:57,  2.02s/it]\n",
      "Loss: 0.01936:  41%|████      | 103/250 [03:31<04:57,  2.02s/it]\n",
      "Loss: 0.02009:  41%|████      | 103/250 [03:31<04:57,  2.02s/it]\n",
      "Loss: 0.02163:  41%|████      | 103/250 [03:31<04:57,  2.02s/it]\n",
      "Loss: 0.02042:  41%|████      | 103/250 [03:31<04:57,  2.02s/it]\n",
      "Loss: 0.02333:  41%|████      | 103/250 [03:31<04:57,  2.02s/it]\n",
      "Loss: 0.02079:  41%|████      | 103/250 [03:32<04:57,  2.02s/it]\n",
      "Loss: 0.02090:  41%|████      | 103/250 [03:32<04:57,  2.02s/it]\n",
      "Loss: 0.02245:  41%|████      | 103/250 [03:32<04:57,  2.02s/it]\n",
      "Loss: 0.02066:  41%|████      | 103/250 [03:32<04:57,  2.02s/it]\n",
      "Loss: 0.01968:  41%|████      | 103/250 [03:32<04:57,  2.02s/it]\n",
      "Loss: 0.02256:  41%|████      | 103/250 [03:32<04:57,  2.02s/it]\n",
      "Loss: 0.02168:  41%|████      | 103/250 [03:32<04:57,  2.02s/it]\n",
      "Loss: 0.01800:  41%|████      | 103/250 [03:32<04:57,  2.02s/it]\n",
      "Loss: 0.01800:  42%|████▏     | 104/250 [03:32<04:55,  2.02s/it]\n",
      "Loss: 0.01983:  42%|████▏     | 104/250 [03:32<04:55,  2.02s/it]\n",
      "Loss: 0.01872:  42%|████▏     | 104/250 [03:32<04:55,  2.02s/it]\n",
      "Loss: 0.01806:  42%|████▏     | 104/250 [03:33<04:55,  2.02s/it]\n",
      "Loss: 0.02212:  42%|████▏     | 104/250 [03:33<04:55,  2.02s/it]\n",
      "Loss: 0.01756:  42%|████▏     | 104/250 [03:33<04:55,  2.02s/it]\n",
      "Loss: 0.02062:  42%|████▏     | 104/250 [03:33<04:55,  2.02s/it]\n",
      "Loss: 0.01954:  42%|████▏     | 104/250 [03:33<04:55,  2.02s/it]\n",
      "Loss: 0.02048:  42%|████▏     | 104/250 [03:33<04:55,  2.02s/it]\n",
      "Loss: 0.02110:  42%|████▏     | 104/250 [03:33<04:55,  2.02s/it]\n",
      "Loss: 0.02037:  42%|████▏     | 104/250 [03:33<04:55,  2.02s/it]\n",
      "Loss: 0.02320:  42%|████▏     | 104/250 [03:33<04:55,  2.02s/it]\n",
      "Loss: 0.02055:  42%|████▏     | 104/250 [03:34<04:55,  2.02s/it]\n",
      "Loss: 0.02099:  42%|████▏     | 104/250 [03:34<04:55,  2.02s/it]\n",
      "Loss: 0.02285:  42%|████▏     | 104/250 [03:34<04:55,  2.02s/it]\n",
      "Loss: 0.02037:  42%|████▏     | 104/250 [03:34<04:55,  2.02s/it]\n",
      "Loss: 0.01898:  42%|████▏     | 104/250 [03:34<04:55,  2.02s/it]\n",
      "Loss: 0.02251:  42%|████▏     | 104/250 [03:34<04:55,  2.02s/it]\n",
      "Loss: 0.02132:  42%|████▏     | 104/250 [03:34<04:55,  2.02s/it]\n",
      "Loss: 0.01752:  42%|████▏     | 104/250 [03:34<04:55,  2.02s/it]\n",
      "Loss: 0.01752:  42%|████▏     | 105/250 [03:34<04:51,  2.01s/it]\n",
      "Loss: 0.01993:  42%|████▏     | 105/250 [03:34<04:51,  2.01s/it]\n",
      "Loss: 0.01889:  42%|████▏     | 105/250 [03:34<04:51,  2.01s/it]\n",
      "Loss: 0.02324:  42%|████▏     | 105/250 [03:35<04:51,  2.01s/it]\n",
      "Loss: 0.01861:  42%|████▏     | 105/250 [03:35<04:51,  2.01s/it]\n",
      "Loss: 0.02022:  42%|████▏     | 105/250 [03:35<04:51,  2.01s/it]\n",
      "Loss: 0.01889:  42%|████▏     | 105/250 [03:35<04:51,  2.01s/it]\n",
      "Loss: 0.02011:  42%|████▏     | 105/250 [03:35<04:51,  2.01s/it]\n",
      "Loss: 0.02109:  42%|████▏     | 105/250 [03:35<04:51,  2.01s/it]\n",
      "Loss: 0.02112:  42%|████▏     | 105/250 [03:35<04:51,  2.01s/it]\n",
      "Loss: 0.02013:  42%|████▏     | 105/250 [03:35<04:51,  2.01s/it]\n",
      "Loss: 0.02278:  42%|████▏     | 105/250 [03:35<04:51,  2.01s/it]\n",
      "Loss: 0.02038:  42%|████▏     | 105/250 [03:36<04:51,  2.01s/it]\n",
      "Loss: 0.02087:  42%|████▏     | 105/250 [03:36<04:51,  2.01s/it]\n",
      "Loss: 0.02237:  42%|████▏     | 105/250 [03:36<04:51,  2.01s/it]\n",
      "Loss: 0.02049:  42%|████▏     | 105/250 [03:36<04:51,  2.01s/it]\n",
      "Loss: 0.01849:  42%|████▏     | 105/250 [03:36<04:51,  2.01s/it]\n",
      "Loss: 0.02243:  42%|████▏     | 105/250 [03:36<04:51,  2.01s/it]\n",
      "Loss: 0.02157:  42%|████▏     | 105/250 [03:36<04:51,  2.01s/it]\n",
      "Loss: 0.01737:  42%|████▏     | 105/250 [03:36<04:51,  2.01s/it]\n",
      "Loss: 0.01737:  42%|████▏     | 106/250 [03:36<04:50,  2.01s/it]\n",
      "Loss: 0.01996:  42%|████▏     | 106/250 [03:36<04:50,  2.01s/it]\n",
      "Loss: 0.02002:  42%|████▏     | 106/250 [03:36<04:50,  2.01s/it]\n",
      "Loss: 0.02270:  42%|████▏     | 106/250 [03:37<04:50,  2.01s/it]\n",
      "Loss: 0.02023:  42%|████▏     | 106/250 [03:37<04:50,  2.01s/it]\n",
      "Loss: 0.01983:  42%|████▏     | 106/250 [03:37<04:50,  2.01s/it]\n",
      "Loss: 0.02065:  42%|████▏     | 106/250 [03:37<04:50,  2.01s/it]\n",
      "Loss: 0.02003:  42%|████▏     | 106/250 [03:37<04:50,  2.01s/it]\n",
      "Loss: 0.02150:  42%|████▏     | 106/250 [03:37<04:50,  2.01s/it]\n",
      "Loss: 0.02050:  42%|████▏     | 106/250 [03:37<04:50,  2.01s/it]\n",
      "Loss: 0.02115:  42%|████▏     | 106/250 [03:37<04:50,  2.01s/it]\n",
      "Loss: 0.02136:  42%|████▏     | 106/250 [03:37<04:50,  2.01s/it]\n",
      "Loss: 0.02179:  42%|████▏     | 106/250 [03:38<04:50,  2.01s/it]\n",
      "Loss: 0.01890:  42%|████▏     | 106/250 [03:38<04:50,  2.01s/it]\n",
      "Loss: 0.01949:  42%|████▏     | 106/250 [03:38<04:50,  2.01s/it]\n",
      "Loss: 0.01874:  42%|████▏     | 106/250 [03:38<04:50,  2.01s/it]\n",
      "Loss: 0.01849:  42%|████▏     | 106/250 [03:38<04:50,  2.01s/it]\n",
      "Loss: 0.02042:  42%|████▏     | 106/250 [03:38<04:50,  2.01s/it]\n",
      "Loss: 0.01783:  42%|████▏     | 106/250 [03:38<04:50,  2.01s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.41it/s]\u001b[A\n",
      "Loss: 0.01783:  43%|████▎     | 107/250 [03:38<04:49,  2.02s/it]\n",
      "Loss: 0.02008:  43%|████▎     | 107/250 [03:38<04:49,  2.02s/it]\n",
      "Loss: 0.01831:  43%|████▎     | 107/250 [03:38<04:49,  2.02s/it]\n",
      "Loss: 0.02259:  43%|████▎     | 107/250 [03:39<04:49,  2.02s/it]\n",
      "Loss: 0.02200:  43%|████▎     | 107/250 [03:39<04:49,  2.02s/it]\n",
      "Loss: 0.02120:  43%|████▎     | 107/250 [03:39<04:49,  2.02s/it]\n",
      "Loss: 0.02075:  43%|████▎     | 107/250 [03:39<04:49,  2.02s/it]\n",
      "Loss: 0.02269:  43%|████▎     | 107/250 [03:39<04:49,  2.02s/it]\n",
      "Loss: 0.02205:  43%|████▎     | 107/250 [03:39<04:49,  2.02s/it]\n",
      "Loss: 0.02002:  43%|████▎     | 107/250 [03:39<04:49,  2.02s/it]\n",
      "Loss: 0.02314:  43%|████▎     | 107/250 [03:39<04:49,  2.02s/it]\n",
      "Loss: 0.02009:  43%|████▎     | 107/250 [03:39<04:49,  2.02s/it]\n",
      "Loss: 0.01866:  43%|████▎     | 107/250 [03:40<04:49,  2.02s/it]\n",
      "Loss: 0.02349:  43%|████▎     | 107/250 [03:40<04:49,  2.02s/it]\n",
      "Loss: 0.02095:  43%|████▎     | 107/250 [03:40<04:49,  2.02s/it]\n",
      "Loss: 0.01947:  43%|████▎     | 107/250 [03:40<04:49,  2.02s/it]\n",
      "Loss: 0.02310:  43%|████▎     | 107/250 [03:40<04:49,  2.02s/it]\n",
      "Loss: 0.01840:  43%|████▎     | 107/250 [03:40<04:49,  2.02s/it]\n",
      "Loss: 0.02084:  43%|████▎     | 107/250 [03:40<04:49,  2.02s/it]\n",
      "Loss: 0.01750:  43%|████▎     | 107/250 [03:40<04:49,  2.02s/it]\n",
      "Loss: 0.01750:  43%|████▎     | 108/250 [03:40<04:48,  2.03s/it]\n",
      "Loss: 0.02036:  43%|████▎     | 108/250 [03:40<04:48,  2.03s/it]\n",
      "Loss: 0.01836:  43%|████▎     | 108/250 [03:41<04:48,  2.03s/it]\n",
      "Loss: 0.02229:  43%|████▎     | 108/250 [03:41<04:48,  2.03s/it]\n",
      "Loss: 0.01907:  43%|████▎     | 108/250 [03:41<04:48,  2.03s/it]\n",
      "Loss: 0.02051:  43%|████▎     | 108/250 [03:41<04:48,  2.03s/it]\n",
      "Loss: 0.01788:  43%|████▎     | 108/250 [03:41<04:48,  2.03s/it]\n",
      "Loss: 0.01947:  43%|████▎     | 108/250 [03:41<04:48,  2.03s/it]\n",
      "Loss: 0.02137:  43%|████▎     | 108/250 [03:41<04:48,  2.03s/it]\n",
      "Loss: 0.02008:  43%|████▎     | 108/250 [03:41<04:48,  2.03s/it]\n",
      "Loss: 0.02093:  43%|████▎     | 108/250 [03:41<04:48,  2.03s/it]\n",
      "Loss: 0.02220:  43%|████▎     | 108/250 [03:42<04:48,  2.03s/it]\n",
      "Loss: 0.02194:  43%|████▎     | 108/250 [03:42<04:48,  2.03s/it]\n",
      "Loss: 0.01908:  43%|████▎     | 108/250 [03:42<04:48,  2.03s/it]\n",
      "Loss: 0.01927:  43%|████▎     | 108/250 [03:42<04:48,  2.03s/it]\n",
      "Loss: 0.01835:  43%|████▎     | 108/250 [03:42<04:48,  2.03s/it]\n",
      "Loss: 0.01840:  43%|████▎     | 108/250 [03:42<04:48,  2.03s/it]\n",
      "Loss: 0.02123:  43%|████▎     | 108/250 [03:42<04:48,  2.03s/it]\n",
      "Loss: 0.01756:  43%|████▎     | 108/250 [03:42<04:48,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 61.95it/s]\u001b[A\n",
      "Loss: 0.01756:  44%|████▎     | 109/250 [03:42<04:45,  2.03s/it]\n",
      "Loss: 0.02018:  44%|████▎     | 109/250 [03:42<04:45,  2.03s/it]\n",
      "Loss: 0.01827:  44%|████▎     | 109/250 [03:43<04:45,  2.03s/it]\n",
      "Loss: 0.02241:  44%|████▎     | 109/250 [03:43<04:45,  2.03s/it]\n",
      "Loss: 0.01872:  44%|████▎     | 109/250 [03:43<04:45,  2.03s/it]\n",
      "Loss: 0.01997:  44%|████▎     | 109/250 [03:43<04:45,  2.03s/it]\n",
      "Loss: 0.01733:  44%|████▎     | 109/250 [03:43<04:45,  2.03s/it]\n",
      "Loss: 0.01938:  44%|████▎     | 109/250 [03:43<04:45,  2.03s/it]\n",
      "Loss: 0.02138:  44%|████▎     | 109/250 [03:43<04:45,  2.03s/it]\n",
      "Loss: 0.02044:  44%|████▎     | 109/250 [03:43<04:45,  2.03s/it]\n",
      "Loss: 0.02098:  44%|████▎     | 109/250 [03:43<04:45,  2.03s/it]\n",
      "Loss: 0.02137:  44%|████▎     | 109/250 [03:44<04:45,  2.03s/it]\n",
      "Loss: 0.02192:  44%|████▎     | 109/250 [03:44<04:45,  2.03s/it]\n",
      "Loss: 0.01889:  44%|████▎     | 109/250 [03:44<04:45,  2.03s/it]\n",
      "Loss: 0.02050:  44%|████▎     | 109/250 [03:44<04:45,  2.03s/it]\n",
      "Loss: 0.01854:  44%|████▎     | 109/250 [03:44<04:45,  2.03s/it]\n",
      "Loss: 0.01804:  44%|████▎     | 109/250 [03:44<04:45,  2.03s/it]\n",
      "Loss: 0.02039:  44%|████▎     | 109/250 [03:44<04:45,  2.03s/it]\n",
      "Loss: 0.01727:  44%|████▎     | 109/250 [03:44<04:45,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 62.64it/s]\u001b[A\n",
      "Loss: 0.01727:  44%|████▍     | 110/250 [03:44<04:42,  2.02s/it]\n",
      "Loss: 0.02008:  44%|████▍     | 110/250 [03:44<04:42,  2.02s/it]\n",
      "Loss: 0.01943:  44%|████▍     | 110/250 [03:45<04:42,  2.02s/it]\n",
      "Loss: 0.02189:  44%|████▍     | 110/250 [03:45<04:42,  2.02s/it]\n",
      "Loss: 0.02034:  44%|████▍     | 110/250 [03:45<04:42,  2.02s/it]\n",
      "Loss: 0.01930:  44%|████▍     | 110/250 [03:45<04:42,  2.02s/it]\n",
      "Loss: 0.01996:  44%|████▍     | 110/250 [03:45<04:42,  2.02s/it]\n",
      "Loss: 0.01991:  44%|████▍     | 110/250 [03:45<04:42,  2.02s/it]\n",
      "Loss: 0.02287:  44%|████▍     | 110/250 [03:45<04:42,  2.02s/it]\n",
      "Loss: 0.02130:  44%|████▍     | 110/250 [03:45<04:42,  2.02s/it]\n",
      "Loss: 0.02141:  44%|████▍     | 110/250 [03:45<04:42,  2.02s/it]\n",
      "Loss: 0.02119:  44%|████▍     | 110/250 [03:46<04:42,  2.02s/it]\n",
      "Loss: 0.02143:  44%|████▍     | 110/250 [03:46<04:42,  2.02s/it]\n",
      "Loss: 0.01859:  44%|████▍     | 110/250 [03:46<04:42,  2.02s/it]\n",
      "Loss: 0.01964:  44%|████▍     | 110/250 [03:46<04:42,  2.02s/it]\n",
      "Loss: 0.01832:  44%|████▍     | 110/250 [03:46<04:42,  2.02s/it]\n",
      "Loss: 0.01791:  44%|████▍     | 110/250 [03:46<04:42,  2.02s/it]\n",
      "Loss: 0.02102:  44%|████▍     | 110/250 [03:46<04:42,  2.02s/it]\n",
      "Loss: 0.01740:  44%|████▍     | 110/250 [03:46<04:42,  2.02s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.47it/s]\u001b[A\n",
      "Loss: 0.01740:  44%|████▍     | 111/250 [03:46<04:41,  2.02s/it]\n",
      "Loss: 0.01984:  44%|████▍     | 111/250 [03:46<04:41,  2.02s/it]\n",
      "Loss: 0.01980:  44%|████▍     | 111/250 [03:47<04:41,  2.02s/it]\n",
      "Loss: 0.02166:  44%|████▍     | 111/250 [03:47<04:41,  2.02s/it]\n",
      "Loss: 0.02033:  44%|████▍     | 111/250 [03:47<04:41,  2.02s/it]\n",
      "Loss: 0.01947:  44%|████▍     | 111/250 [03:47<04:41,  2.02s/it]\n",
      "Loss: 0.01997:  44%|████▍     | 111/250 [03:47<04:41,  2.02s/it]\n",
      "Loss: 0.01971:  44%|████▍     | 111/250 [03:47<04:41,  2.02s/it]\n",
      "Loss: 0.02021:  44%|████▍     | 111/250 [03:47<04:41,  2.02s/it]\n",
      "Loss: 0.01883:  44%|████▍     | 111/250 [03:47<04:41,  2.02s/it]\n",
      "Loss: 0.02094:  44%|████▍     | 111/250 [03:47<04:41,  2.02s/it]\n",
      "Loss: 0.02202:  44%|████▍     | 111/250 [03:48<04:41,  2.02s/it]\n",
      "Loss: 0.02014:  44%|████▍     | 111/250 [03:48<04:41,  2.02s/it]\n",
      "Loss: 0.01999:  44%|████▍     | 111/250 [03:48<04:41,  2.02s/it]\n",
      "Loss: 0.01935:  44%|████▍     | 111/250 [03:48<04:41,  2.02s/it]\n",
      "Loss: 0.01963:  44%|████▍     | 111/250 [03:48<04:41,  2.02s/it]\n",
      "Loss: 0.02171:  44%|████▍     | 111/250 [03:48<04:41,  2.02s/it]\n",
      "Loss: 0.01968:  44%|████▍     | 111/250 [03:48<04:41,  2.02s/it]\n",
      "Loss: 0.01776:  44%|████▍     | 111/250 [03:48<04:41,  2.02s/it]\n",
      "Loss: 0.01776:  45%|████▍     | 112/250 [03:48<04:38,  2.02s/it]\n",
      "Loss: 0.02002:  45%|████▍     | 112/250 [03:48<04:38,  2.02s/it]\n",
      "Loss: 0.01852:  45%|████▍     | 112/250 [03:49<04:38,  2.02s/it]\n",
      "Loss: 0.01756:  45%|████▍     | 112/250 [03:49<04:38,  2.02s/it]\n",
      "Loss: 0.02122:  45%|████▍     | 112/250 [03:49<04:38,  2.02s/it]\n",
      "Loss: 0.01691:  45%|████▍     | 112/250 [03:49<04:38,  2.02s/it]\n",
      "Loss: 0.02069:  45%|████▍     | 112/250 [03:49<04:38,  2.02s/it]\n",
      "Loss: 0.02170:  45%|████▍     | 112/250 [03:49<04:38,  2.02s/it]\n",
      "Loss: 0.02051:  45%|████▍     | 112/250 [03:49<04:38,  2.02s/it]\n",
      "Loss: 0.02124:  45%|████▍     | 112/250 [03:49<04:38,  2.02s/it]\n",
      "Loss: 0.01986:  45%|████▍     | 112/250 [03:49<04:38,  2.02s/it]\n",
      "Loss: 0.01856:  45%|████▍     | 112/250 [03:50<04:38,  2.02s/it]\n",
      "Loss: 0.01945:  45%|████▍     | 112/250 [03:50<04:38,  2.02s/it]\n",
      "Loss: 0.02019:  45%|████▍     | 112/250 [03:50<04:38,  2.02s/it]\n",
      "Loss: 0.01771:  45%|████▍     | 112/250 [03:50<04:38,  2.02s/it]\n",
      "Loss: 0.02083:  45%|████▍     | 112/250 [03:50<04:38,  2.02s/it]\n",
      "Loss: 0.02070:  45%|████▍     | 112/250 [03:50<04:38,  2.02s/it]\n",
      "Loss: 0.02255:  45%|████▍     | 112/250 [03:50<04:38,  2.02s/it]\n",
      "Loss: 0.01873:  45%|████▍     | 112/250 [03:50<04:38,  2.02s/it]\n",
      "Loss: 0.01749:  45%|████▍     | 112/250 [03:50<04:38,  2.02s/it]\n",
      "Loss: 0.01749:  45%|████▌     | 113/250 [03:50<04:37,  2.02s/it]\n",
      "Loss: 0.02012:  45%|████▌     | 113/250 [03:51<04:37,  2.02s/it]\n",
      "Loss: 0.01921:  45%|████▌     | 113/250 [03:51<04:37,  2.02s/it]\n",
      "Loss: 0.02158:  45%|████▌     | 113/250 [03:51<04:37,  2.02s/it]\n",
      "Loss: 0.01896:  45%|████▌     | 113/250 [03:51<04:37,  2.02s/it]\n",
      "Loss: 0.01697:  45%|████▌     | 113/250 [03:51<04:37,  2.02s/it]\n",
      "Loss: 0.02027:  45%|████▌     | 113/250 [03:51<04:37,  2.02s/it]\n",
      "Loss: 0.01821:  45%|████▌     | 113/250 [03:51<04:37,  2.02s/it]\n",
      "Loss: 0.02032:  45%|████▌     | 113/250 [03:51<04:37,  2.02s/it]\n",
      "Loss: 0.02165:  45%|████▌     | 113/250 [03:51<04:37,  2.02s/it]\n",
      "Loss: 0.02160:  45%|████▌     | 113/250 [03:51<04:37,  2.02s/it]\n",
      "Loss: 0.01959:  45%|████▌     | 113/250 [03:52<04:37,  2.02s/it]\n",
      "Loss: 0.01886:  45%|████▌     | 113/250 [03:52<04:37,  2.02s/it]\n",
      "Loss: 0.02321:  45%|████▌     | 113/250 [03:52<04:37,  2.02s/it]\n",
      "Loss: 0.02071:  45%|████▌     | 113/250 [03:52<04:37,  2.02s/it]\n",
      "Loss: 0.01945:  45%|████▌     | 113/250 [03:52<04:37,  2.02s/it]\n",
      "Loss: 0.02246:  45%|████▌     | 113/250 [03:52<04:37,  2.02s/it]\n",
      "Loss: 0.01821:  45%|████▌     | 113/250 [03:52<04:37,  2.02s/it]\n",
      "Loss: 0.01989:  45%|████▌     | 113/250 [03:52<04:37,  2.02s/it]\n",
      "Loss: 0.01700:  45%|████▌     | 113/250 [03:52<04:37,  2.02s/it]\n",
      "Loss: 0.01700:  46%|████▌     | 114/250 [03:52<04:35,  2.02s/it]\n",
      "Loss: 0.01962:  46%|████▌     | 114/250 [03:53<04:35,  2.02s/it]\n",
      "Loss: 0.01947:  46%|████▌     | 114/250 [03:53<04:35,  2.02s/it]\n",
      "Loss: 0.02173:  46%|████▌     | 114/250 [03:53<04:35,  2.02s/it]\n",
      "Loss: 0.01770:  46%|████▌     | 114/250 [03:53<04:35,  2.02s/it]\n",
      "Loss: 0.01980:  46%|████▌     | 114/250 [03:53<04:35,  2.02s/it]\n",
      "Loss: 0.01720:  46%|████▌     | 114/250 [03:53<04:35,  2.02s/it]\n",
      "Loss: 0.01971:  46%|████▌     | 114/250 [03:53<04:35,  2.02s/it]\n",
      "Loss: 0.02146:  46%|████▌     | 114/250 [03:53<04:35,  2.02s/it]\n",
      "Loss: 0.01961:  46%|████▌     | 114/250 [03:53<04:35,  2.02s/it]\n",
      "Loss: 0.02044:  46%|████▌     | 114/250 [03:54<04:35,  2.02s/it]\n",
      "Loss: 0.02054:  46%|████▌     | 114/250 [03:54<04:35,  2.02s/it]\n",
      "Loss: 0.02159:  46%|████▌     | 114/250 [03:54<04:35,  2.02s/it]\n",
      "Loss: 0.01838:  46%|████▌     | 114/250 [03:54<04:35,  2.02s/it]\n",
      "Loss: 0.01891:  46%|████▌     | 114/250 [03:54<04:35,  2.02s/it]\n",
      "Loss: 0.01787:  46%|████▌     | 114/250 [03:54<04:35,  2.02s/it]\n",
      "Loss: 0.01818:  46%|████▌     | 114/250 [03:54<04:35,  2.02s/it]\n",
      "Loss: 0.01977:  46%|████▌     | 114/250 [03:54<04:35,  2.02s/it]\n",
      "Loss: 0.01693:  46%|████▌     | 114/250 [03:54<04:35,  2.02s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 60.25it/s]\u001b[A\n",
      "Loss: 0.01693:  46%|████▌     | 115/250 [03:54<04:33,  2.02s/it]\n",
      "Loss: 0.01946:  46%|████▌     | 115/250 [03:55<04:33,  2.02s/it]\n",
      "Loss: 0.01938:  46%|████▌     | 115/250 [03:55<04:33,  2.02s/it]\n",
      "Loss: 0.02133:  46%|████▌     | 115/250 [03:55<04:33,  2.02s/it]\n",
      "Loss: 0.01801:  46%|████▌     | 115/250 [03:55<04:33,  2.02s/it]\n",
      "Loss: 0.01965:  46%|████▌     | 115/250 [03:55<04:33,  2.02s/it]\n",
      "Loss: 0.01707:  46%|████▌     | 115/250 [03:55<04:33,  2.02s/it]\n",
      "Loss: 0.02003:  46%|████▌     | 115/250 [03:55<04:33,  2.02s/it]\n",
      "Loss: 0.02114:  46%|████▌     | 115/250 [03:55<04:33,  2.02s/it]\n",
      "Loss: 0.01945:  46%|████▌     | 115/250 [03:55<04:33,  2.02s/it]\n",
      "Loss: 0.02087:  46%|████▌     | 115/250 [03:56<04:33,  2.02s/it]\n",
      "Loss: 0.02034:  46%|████▌     | 115/250 [03:56<04:33,  2.02s/it]\n",
      "Loss: 0.02073:  46%|████▌     | 115/250 [03:56<04:33,  2.02s/it]\n",
      "Loss: 0.01874:  46%|████▌     | 115/250 [03:56<04:33,  2.02s/it]\n",
      "Loss: 0.01873:  46%|████▌     | 115/250 [03:56<04:33,  2.02s/it]\n",
      "Loss: 0.01780:  46%|████▌     | 115/250 [03:56<04:33,  2.02s/it]\n",
      "Loss: 0.01781:  46%|████▌     | 115/250 [03:56<04:33,  2.02s/it]\n",
      "Loss: 0.01951:  46%|████▌     | 115/250 [03:56<04:33,  2.02s/it]\n",
      "Loss: 0.01681:  46%|████▌     | 115/250 [03:56<04:33,  2.02s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 62.47it/s]\u001b[A\n",
      "Loss: 0.01681:  46%|████▋     | 116/250 [03:56<04:31,  2.02s/it]\n",
      "Loss: 0.01944:  46%|████▋     | 116/250 [03:57<04:31,  2.02s/it]\n",
      "Loss: 0.01914:  46%|████▋     | 116/250 [03:57<04:31,  2.02s/it]\n",
      "Loss: 0.02159:  46%|████▋     | 116/250 [03:57<04:31,  2.02s/it]\n",
      "Loss: 0.01981:  46%|████▋     | 116/250 [03:57<04:31,  2.02s/it]\n",
      "Loss: 0.01944:  46%|████▋     | 116/250 [03:57<04:31,  2.02s/it]\n",
      "Loss: 0.01989:  46%|████▋     | 116/250 [03:57<04:31,  2.02s/it]\n",
      "Loss: 0.01901:  46%|████▋     | 116/250 [03:57<04:31,  2.02s/it]\n",
      "Loss: 0.02029:  46%|████▋     | 116/250 [03:57<04:31,  2.02s/it]\n",
      "Loss: 0.02098:  46%|████▋     | 116/250 [03:57<04:31,  2.02s/it]\n",
      "Loss: 0.01911:  46%|████▋     | 116/250 [03:58<04:31,  2.02s/it]\n",
      "Loss: 0.01921:  46%|████▋     | 116/250 [03:58<04:31,  2.02s/it]\n",
      "Loss: 0.01840:  46%|████▋     | 116/250 [03:58<04:31,  2.02s/it]\n",
      "Loss: 0.02318:  46%|████▋     | 116/250 [03:58<04:31,  2.02s/it]\n",
      "Loss: 0.02017:  46%|████▋     | 116/250 [03:58<04:31,  2.02s/it]\n",
      "Loss: 0.01891:  46%|████▋     | 116/250 [03:58<04:31,  2.02s/it]\n",
      "Loss: 0.02270:  46%|████▋     | 116/250 [03:58<04:31,  2.02s/it]\n",
      "Loss: 0.01807:  46%|████▋     | 116/250 [03:58<04:31,  2.02s/it]\n",
      "Loss: 0.01942:  46%|████▋     | 116/250 [03:58<04:31,  2.02s/it]\n",
      "Loss: 0.01708:  46%|████▋     | 116/250 [03:59<04:31,  2.02s/it]\n",
      "Loss: 0.01708:  47%|████▋     | 117/250 [03:59<04:30,  2.03s/it]\n",
      "Loss: 0.01989:  47%|████▋     | 117/250 [03:59<04:30,  2.03s/it]\n",
      "Loss: 0.01912:  47%|████▋     | 117/250 [03:59<04:30,  2.03s/it]\n",
      "Loss: 0.02154:  47%|████▋     | 117/250 [03:59<04:30,  2.03s/it]\n",
      "Loss: 0.02092:  47%|████▋     | 117/250 [03:59<04:30,  2.03s/it]\n",
      "Loss: 0.01699:  47%|████▋     | 117/250 [03:59<04:30,  2.03s/it]\n",
      "Loss: 0.02106:  47%|████▋     | 117/250 [03:59<04:30,  2.03s/it]\n",
      "Loss: 0.02179:  47%|████▋     | 117/250 [03:59<04:30,  2.03s/it]\n",
      "Loss: 0.01980:  47%|████▋     | 117/250 [03:59<04:30,  2.03s/it]\n",
      "Loss: 0.02072:  47%|████▋     | 117/250 [04:00<04:30,  2.03s/it]\n",
      "Loss: 0.01955:  47%|████▋     | 117/250 [04:00<04:30,  2.03s/it]\n",
      "Loss: 0.01895:  47%|████▋     | 117/250 [04:00<04:30,  2.03s/it]\n",
      "Loss: 0.01897:  47%|████▋     | 117/250 [04:00<04:30,  2.03s/it]\n",
      "Loss: 0.01998:  47%|████▋     | 117/250 [04:00<04:30,  2.03s/it]\n",
      "Loss: 0.01760:  47%|████▋     | 117/250 [04:00<04:30,  2.03s/it]\n",
      "Loss: 0.02016:  47%|████▋     | 117/250 [04:00<04:30,  2.03s/it]\n",
      "Loss: 0.02089:  47%|████▋     | 117/250 [04:00<04:30,  2.03s/it]\n",
      "Loss: 0.02191:  47%|████▋     | 117/250 [04:00<04:30,  2.03s/it]\n",
      "Loss: 0.01836:  47%|████▋     | 117/250 [04:01<04:30,  2.03s/it]\n",
      "Loss: 0.01730:  47%|████▋     | 117/250 [04:01<04:30,  2.03s/it]\n",
      "Loss: 0.01730:  47%|████▋     | 118/250 [04:01<04:29,  2.04s/it]\n",
      "Loss: 0.01954:  47%|████▋     | 118/250 [04:01<04:29,  2.04s/it]\n",
      "Loss: 0.01951:  47%|████▋     | 118/250 [04:01<04:29,  2.04s/it]\n",
      "Loss: 0.02164:  47%|████▋     | 118/250 [04:01<04:29,  2.04s/it]\n",
      "Loss: 0.01775:  47%|████▋     | 118/250 [04:01<04:29,  2.04s/it]\n",
      "Loss: 0.01945:  47%|████▋     | 118/250 [04:01<04:29,  2.04s/it]\n",
      "Loss: 0.01770:  47%|████▋     | 118/250 [04:01<04:29,  2.04s/it]\n",
      "Loss: 0.01885:  47%|████▋     | 118/250 [04:01<04:29,  2.04s/it]\n",
      "Loss: 0.02074:  47%|████▋     | 118/250 [04:01<04:29,  2.04s/it]\n",
      "Loss: 0.02018:  47%|████▋     | 118/250 [04:02<04:29,  2.04s/it]\n",
      "Loss: 0.02060:  47%|████▋     | 118/250 [04:02<04:29,  2.04s/it]\n",
      "Loss: 0.02037:  47%|████▋     | 118/250 [04:02<04:29,  2.04s/it]\n",
      "Loss: 0.02109:  47%|████▋     | 118/250 [04:02<04:29,  2.04s/it]\n",
      "Loss: 0.01853:  47%|████▋     | 118/250 [04:02<04:29,  2.04s/it]\n",
      "Loss: 0.01926:  47%|████▋     | 118/250 [04:02<04:29,  2.04s/it]\n",
      "Loss: 0.01794:  47%|████▋     | 118/250 [04:02<04:29,  2.04s/it]\n",
      "Loss: 0.01777:  47%|████▋     | 118/250 [04:02<04:29,  2.04s/it]\n",
      "Loss: 0.01982:  47%|████▋     | 118/250 [04:02<04:29,  2.04s/it]\n",
      "Loss: 0.01701:  47%|████▋     | 118/250 [04:03<04:29,  2.04s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 64.34it/s]\u001b[A\n",
      "Loss: 0.01701:  48%|████▊     | 119/250 [04:03<04:24,  2.02s/it]\n",
      "Loss: 0.01929:  48%|████▊     | 119/250 [04:03<04:24,  2.02s/it]\n",
      "Loss: 0.01931:  48%|████▊     | 119/250 [04:03<04:24,  2.02s/it]\n",
      "Loss: 0.02167:  48%|████▊     | 119/250 [04:03<04:24,  2.02s/it]\n",
      "Loss: 0.01763:  48%|████▊     | 119/250 [04:03<04:24,  2.02s/it]\n",
      "Loss: 0.01895:  48%|████▊     | 119/250 [04:03<04:24,  2.02s/it]\n",
      "Loss: 0.01672:  48%|████▊     | 119/250 [04:03<04:24,  2.02s/it]\n",
      "Loss: 0.01932:  48%|████▊     | 119/250 [04:03<04:24,  2.02s/it]\n",
      "Loss: 0.02103:  48%|████▊     | 119/250 [04:03<04:24,  2.02s/it]\n",
      "Loss: 0.02043:  48%|████▊     | 119/250 [04:04<04:24,  2.02s/it]\n",
      "Loss: 0.01916:  48%|████▊     | 119/250 [04:04<04:24,  2.02s/it]\n",
      "Loss: 0.02193:  48%|████▊     | 119/250 [04:04<04:24,  2.02s/it]\n",
      "Loss: 0.01983:  48%|████▊     | 119/250 [04:04<04:24,  2.02s/it]\n",
      "Loss: 0.02023:  48%|████▊     | 119/250 [04:04<04:24,  2.02s/it]\n",
      "Loss: 0.02104:  48%|████▊     | 119/250 [04:04<04:24,  2.02s/it]\n",
      "Loss: 0.02037:  48%|████▊     | 119/250 [04:04<04:24,  2.02s/it]\n",
      "Loss: 0.01756:  48%|████▊     | 119/250 [04:04<04:24,  2.02s/it]\n",
      "Loss: 0.02087:  48%|████▊     | 119/250 [04:04<04:24,  2.02s/it]\n",
      "Loss: 0.02105:  48%|████▊     | 119/250 [04:05<04:24,  2.02s/it]\n",
      "Loss: 0.01767:  48%|████▊     | 119/250 [04:05<04:24,  2.02s/it]\n",
      "Loss: 0.01767:  48%|████▊     | 120/250 [04:05<04:23,  2.02s/it]\n",
      "Loss: 0.01972:  48%|████▊     | 120/250 [04:05<04:23,  2.02s/it]\n",
      "Loss: 0.01953:  48%|████▊     | 120/250 [04:05<04:23,  2.02s/it]\n",
      "Loss: 0.02113:  48%|████▊     | 120/250 [04:05<04:23,  2.02s/it]\n",
      "Loss: 0.01941:  48%|████▊     | 120/250 [04:05<04:23,  2.02s/it]\n",
      "Loss: 0.01939:  48%|████▊     | 120/250 [04:05<04:23,  2.02s/it]\n",
      "Loss: 0.01894:  48%|████▊     | 120/250 [04:05<04:23,  2.02s/it]\n",
      "Loss: 0.01940:  48%|████▊     | 120/250 [04:05<04:23,  2.02s/it]\n",
      "Loss: 0.01948:  48%|████▊     | 120/250 [04:05<04:23,  2.02s/it]\n",
      "Loss: 0.01779:  48%|████▊     | 120/250 [04:06<04:23,  2.02s/it]\n",
      "Loss: 0.02101:  48%|████▊     | 120/250 [04:06<04:23,  2.02s/it]\n",
      "Loss: 0.02177:  48%|████▊     | 120/250 [04:06<04:23,  2.02s/it]\n",
      "Loss: 0.02046:  48%|████▊     | 120/250 [04:06<04:23,  2.02s/it]\n",
      "Loss: 0.01919:  48%|████▊     | 120/250 [04:06<04:23,  2.02s/it]\n",
      "Loss: 0.01920:  48%|████▊     | 120/250 [04:06<04:23,  2.02s/it]\n",
      "Loss: 0.01882:  48%|████▊     | 120/250 [04:06<04:23,  2.02s/it]\n",
      "Loss: 0.02119:  48%|████▊     | 120/250 [04:06<04:23,  2.02s/it]\n",
      "Loss: 0.01915:  48%|████▊     | 120/250 [04:06<04:23,  2.02s/it]\n",
      "Loss: 0.01665:  48%|████▊     | 120/250 [04:07<04:23,  2.02s/it]\n",
      "Loss: 0.01665:  48%|████▊     | 121/250 [04:07<04:19,  2.01s/it]\n",
      "Loss: 0.01983:  48%|████▊     | 121/250 [04:07<04:19,  2.01s/it]\n",
      "Loss: 0.01905:  48%|████▊     | 121/250 [04:07<04:19,  2.01s/it]\n",
      "Loss: 0.02134:  48%|████▊     | 121/250 [04:07<04:19,  2.01s/it]\n",
      "Loss: 0.01820:  48%|████▊     | 121/250 [04:07<04:19,  2.01s/it]\n",
      "Loss: 0.01876:  48%|████▊     | 121/250 [04:07<04:19,  2.01s/it]\n",
      "Loss: 0.01958:  48%|████▊     | 121/250 [04:07<04:19,  2.01s/it]\n",
      "Loss: 0.02122:  48%|████▊     | 121/250 [04:07<04:19,  2.01s/it]\n",
      "Loss: 0.01936:  48%|████▊     | 121/250 [04:07<04:19,  2.01s/it]\n",
      "Loss: 0.02036:  48%|████▊     | 121/250 [04:08<04:19,  2.01s/it]\n",
      "Loss: 0.01976:  48%|████▊     | 121/250 [04:08<04:19,  2.01s/it]\n",
      "Loss: 0.01878:  48%|████▊     | 121/250 [04:08<04:19,  2.01s/it]\n",
      "Loss: 0.01885:  48%|████▊     | 121/250 [04:08<04:19,  2.01s/it]\n",
      "Loss: 0.01972:  48%|████▊     | 121/250 [04:08<04:19,  2.01s/it]\n",
      "Loss: 0.01750:  48%|████▊     | 121/250 [04:08<04:19,  2.01s/it]\n",
      "Loss: 0.02058:  48%|████▊     | 121/250 [04:08<04:19,  2.01s/it]\n",
      "Loss: 0.02113:  48%|████▊     | 121/250 [04:08<04:19,  2.01s/it]\n",
      "Loss: 0.02135:  48%|████▊     | 121/250 [04:08<04:19,  2.01s/it]\n",
      "Loss: 0.01804:  48%|████▊     | 121/250 [04:09<04:19,  2.01s/it]\n",
      "Loss: 0.01671:  48%|████▊     | 121/250 [04:09<04:19,  2.01s/it]\n",
      "Loss: 0.01671:  49%|████▉     | 122/250 [04:09<04:18,  2.02s/it]\n",
      "Loss: 0.01968:  49%|████▉     | 122/250 [04:09<04:18,  2.02s/it]\n",
      "Loss: 0.01907:  49%|████▉     | 122/250 [04:09<04:18,  2.02s/it]\n",
      "Loss: 0.02105:  49%|████▉     | 122/250 [04:09<04:18,  2.02s/it]\n",
      "Loss: 0.01961:  49%|████▉     | 122/250 [04:09<04:18,  2.02s/it]\n",
      "Loss: 0.01823:  49%|████▉     | 122/250 [04:09<04:18,  2.02s/it]\n",
      "Loss: 0.01860:  49%|████▉     | 122/250 [04:09<04:18,  2.02s/it]\n",
      "Loss: 0.01958:  49%|████▉     | 122/250 [04:09<04:18,  2.02s/it]\n",
      "Loss: 0.02026:  49%|████▉     | 122/250 [04:10<04:18,  2.02s/it]\n",
      "Loss: 0.01963:  49%|████▉     | 122/250 [04:10<04:18,  2.02s/it]\n",
      "Loss: 0.02063:  49%|████▉     | 122/250 [04:10<04:18,  2.02s/it]\n",
      "Loss: 0.02001:  49%|████▉     | 122/250 [04:10<04:18,  2.02s/it]\n",
      "Loss: 0.02058:  49%|████▉     | 122/250 [04:10<04:18,  2.02s/it]\n",
      "Loss: 0.01875:  49%|████▉     | 122/250 [04:10<04:18,  2.02s/it]\n",
      "Loss: 0.01909:  49%|████▉     | 122/250 [04:10<04:18,  2.02s/it]\n",
      "Loss: 0.01778:  49%|████▉     | 122/250 [04:10<04:18,  2.02s/it]\n",
      "Loss: 0.01786:  49%|████▉     | 122/250 [04:10<04:18,  2.02s/it]\n",
      "Loss: 0.01944:  49%|████▉     | 122/250 [04:11<04:18,  2.02s/it]\n",
      "Loss: 0.01665:  49%|████▉     | 122/250 [04:11<04:18,  2.02s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 60.40it/s]\u001b[A\n",
      "Loss: 0.01665:  49%|████▉     | 123/250 [04:11<04:17,  2.03s/it]\n",
      "Loss: 0.01931:  49%|████▉     | 123/250 [04:11<04:17,  2.03s/it]\n",
      "Loss: 0.01895:  49%|████▉     | 123/250 [04:11<04:17,  2.03s/it]\n",
      "Loss: 0.02115:  49%|████▉     | 123/250 [04:11<04:17,  2.03s/it]\n",
      "Loss: 0.01952:  49%|████▉     | 123/250 [04:11<04:17,  2.03s/it]\n",
      "Loss: 0.01904:  49%|████▉     | 123/250 [04:11<04:17,  2.03s/it]\n",
      "Loss: 0.01765:  49%|████▉     | 123/250 [04:11<04:17,  2.03s/it]\n",
      "Loss: 0.01869:  49%|████▉     | 123/250 [04:11<04:17,  2.03s/it]\n",
      "Loss: 0.02050:  49%|████▉     | 123/250 [04:12<04:17,  2.03s/it]\n",
      "Loss: 0.02016:  49%|████▉     | 123/250 [04:12<04:17,  2.03s/it]\n",
      "Loss: 0.01931:  49%|████▉     | 123/250 [04:12<04:17,  2.03s/it]\n",
      "Loss: 0.01853:  49%|████▉     | 123/250 [04:12<04:17,  2.03s/it]\n",
      "Loss: 0.01889:  49%|████▉     | 123/250 [04:12<04:17,  2.03s/it]\n",
      "Loss: 0.01948:  49%|████▉     | 123/250 [04:12<04:17,  2.03s/it]\n",
      "Loss: 0.01779:  49%|████▉     | 123/250 [04:12<04:17,  2.03s/it]\n",
      "Loss: 0.01992:  49%|████▉     | 123/250 [04:12<04:17,  2.03s/it]\n",
      "Loss: 0.02021:  49%|████▉     | 123/250 [04:12<04:17,  2.03s/it]\n",
      "Loss: 0.02141:  49%|████▉     | 123/250 [04:13<04:17,  2.03s/it]\n",
      "Loss: 0.01829:  49%|████▉     | 123/250 [04:13<04:17,  2.03s/it]\n",
      "Loss: 0.01691:  49%|████▉     | 123/250 [04:13<04:17,  2.03s/it]\n",
      "Loss: 0.01691:  50%|████▉     | 124/250 [04:13<04:16,  2.03s/it]\n",
      "Loss: 0.01913:  50%|████▉     | 124/250 [04:13<04:16,  2.03s/it]\n",
      "Loss: 0.01900:  50%|████▉     | 124/250 [04:13<04:16,  2.03s/it]\n",
      "Loss: 0.02092:  50%|████▉     | 124/250 [04:13<04:16,  2.03s/it]\n",
      "Loss: 0.01799:  50%|████▉     | 124/250 [04:13<04:16,  2.03s/it]\n",
      "Loss: 0.01878:  50%|████▉     | 124/250 [04:13<04:16,  2.03s/it]\n",
      "Loss: 0.01639:  50%|████▉     | 124/250 [04:13<04:16,  2.03s/it]\n",
      "Loss: 0.01866:  50%|████▉     | 124/250 [04:13<04:16,  2.03s/it]\n",
      "Loss: 0.01919:  50%|████▉     | 124/250 [04:14<04:16,  2.03s/it]\n",
      "Loss: 0.02021:  50%|████▉     | 124/250 [04:14<04:16,  2.03s/it]\n",
      "Loss: 0.01905:  50%|████▉     | 124/250 [04:14<04:16,  2.03s/it]\n",
      "Loss: 0.01851:  50%|████▉     | 124/250 [04:14<04:16,  2.03s/it]\n",
      "Loss: 0.01895:  50%|████▉     | 124/250 [04:14<04:16,  2.03s/it]\n",
      "Loss: 0.02000:  50%|████▉     | 124/250 [04:14<04:16,  2.03s/it]\n",
      "Loss: 0.01712:  50%|████▉     | 124/250 [04:14<04:16,  2.03s/it]\n",
      "Loss: 0.02003:  50%|████▉     | 124/250 [04:14<04:16,  2.03s/it]\n",
      "Loss: 0.02042:  50%|████▉     | 124/250 [04:14<04:16,  2.03s/it]\n",
      "Loss: 0.02168:  50%|████▉     | 124/250 [04:15<04:16,  2.03s/it]\n",
      "Loss: 0.01799:  50%|████▉     | 124/250 [04:15<04:16,  2.03s/it]\n",
      "Loss: 0.01677:  50%|████▉     | 124/250 [04:15<04:16,  2.03s/it]\n",
      "Loss: 0.01677:  50%|█████     | 125/250 [04:15<04:14,  2.04s/it]\n",
      "Loss: 0.01908:  50%|█████     | 125/250 [04:15<04:14,  2.04s/it]\n",
      "Loss: 0.01778:  50%|█████     | 125/250 [04:15<04:14,  2.04s/it]\n",
      "Loss: 0.01771:  50%|█████     | 125/250 [04:15<04:14,  2.04s/it]\n",
      "Loss: 0.02116:  50%|█████     | 125/250 [04:15<04:14,  2.04s/it]\n",
      "Loss: 0.01651:  50%|█████     | 125/250 [04:15<04:14,  2.04s/it]\n",
      "Loss: 0.02015:  50%|█████     | 125/250 [04:15<04:14,  2.04s/it]\n",
      "Loss: 0.01833:  50%|█████     | 125/250 [04:16<04:14,  2.04s/it]\n",
      "Loss: 0.01882:  50%|█████     | 125/250 [04:16<04:14,  2.04s/it]\n",
      "Loss: 0.02032:  50%|█████     | 125/250 [04:16<04:14,  2.04s/it]\n",
      "Loss: 0.01873:  50%|█████     | 125/250 [04:16<04:14,  2.04s/it]\n",
      "Loss: 0.02191:  50%|█████     | 125/250 [04:16<04:14,  2.04s/it]\n",
      "Loss: 0.01964:  50%|█████     | 125/250 [04:16<04:14,  2.04s/it]\n",
      "Loss: 0.02012:  50%|█████     | 125/250 [04:16<04:14,  2.04s/it]\n",
      "Loss: 0.02109:  50%|█████     | 125/250 [04:16<04:14,  2.04s/it]\n",
      "Loss: 0.01965:  50%|█████     | 125/250 [04:16<04:14,  2.04s/it]\n",
      "Loss: 0.01859:  50%|█████     | 125/250 [04:17<04:14,  2.04s/it]\n",
      "Loss: 0.02049:  50%|█████     | 125/250 [04:17<04:14,  2.04s/it]\n",
      "Loss: 0.02031:  50%|█████     | 125/250 [04:17<04:14,  2.04s/it]\n",
      "Loss: 0.01663:  50%|█████     | 125/250 [04:17<04:14,  2.04s/it]\n",
      "Loss: 0.01663:  50%|█████     | 126/250 [04:17<04:12,  2.04s/it]\n",
      "Loss: 0.01950:  50%|█████     | 126/250 [04:17<04:12,  2.04s/it]\n",
      "Loss: 0.01958:  50%|█████     | 126/250 [04:17<04:12,  2.04s/it]\n",
      "Loss: 0.02146:  50%|█████     | 126/250 [04:17<04:12,  2.04s/it]\n",
      "Loss: 0.01759:  50%|█████     | 126/250 [04:17<04:12,  2.04s/it]\n",
      "Loss: 0.01908:  50%|█████     | 126/250 [04:17<04:12,  2.04s/it]\n",
      "Loss: 0.01937:  50%|█████     | 126/250 [04:17<04:12,  2.04s/it]\n",
      "Loss: 0.01815:  50%|█████     | 126/250 [04:18<04:12,  2.04s/it]\n",
      "Loss: 0.01887:  50%|█████     | 126/250 [04:18<04:12,  2.04s/it]\n",
      "Loss: 0.01992:  50%|█████     | 126/250 [04:18<04:12,  2.04s/it]\n",
      "Loss: 0.01891:  50%|█████     | 126/250 [04:18<04:12,  2.04s/it]\n",
      "Loss: 0.02192:  50%|█████     | 126/250 [04:18<04:12,  2.04s/it]\n",
      "Loss: 0.01913:  50%|█████     | 126/250 [04:18<04:12,  2.04s/it]\n",
      "Loss: 0.01941:  50%|█████     | 126/250 [04:18<04:12,  2.04s/it]\n",
      "Loss: 0.02110:  50%|█████     | 126/250 [04:18<04:12,  2.04s/it]\n",
      "Loss: 0.01954:  50%|█████     | 126/250 [04:18<04:12,  2.04s/it]\n",
      "Loss: 0.01825:  50%|█████     | 126/250 [04:19<04:12,  2.04s/it]\n",
      "Loss: 0.02019:  50%|█████     | 126/250 [04:19<04:12,  2.04s/it]\n",
      "Loss: 0.02041:  50%|█████     | 126/250 [04:19<04:12,  2.04s/it]\n",
      "Loss: 0.01660:  50%|█████     | 126/250 [04:19<04:12,  2.04s/it]\n",
      "Loss: 0.01660:  51%|█████     | 127/250 [04:19<04:10,  2.03s/it]\n",
      "Loss: 0.01921:  51%|█████     | 127/250 [04:19<04:10,  2.03s/it]\n",
      "Loss: 0.01862:  51%|█████     | 127/250 [04:19<04:10,  2.03s/it]\n",
      "Loss: 0.02089:  51%|█████     | 127/250 [04:19<04:10,  2.03s/it]\n",
      "Loss: 0.01740:  51%|█████     | 127/250 [04:19<04:10,  2.03s/it]\n",
      "Loss: 0.01867:  51%|█████     | 127/250 [04:19<04:10,  2.03s/it]\n",
      "Loss: 0.01942:  51%|█████     | 127/250 [04:19<04:10,  2.03s/it]\n",
      "Loss: 0.01843:  51%|█████     | 127/250 [04:20<04:10,  2.03s/it]\n",
      "Loss: 0.01903:  51%|█████     | 127/250 [04:20<04:10,  2.03s/it]\n",
      "Loss: 0.02027:  51%|█████     | 127/250 [04:20<04:10,  2.03s/it]\n",
      "Loss: 0.01901:  51%|█████     | 127/250 [04:20<04:10,  2.03s/it]\n",
      "Loss: 0.01855:  51%|█████     | 127/250 [04:20<04:10,  2.03s/it]\n",
      "Loss: 0.01763:  51%|█████     | 127/250 [04:20<04:10,  2.03s/it]\n",
      "Loss: 0.02256:  51%|█████     | 127/250 [04:20<04:10,  2.03s/it]\n",
      "Loss: 0.02003:  51%|█████     | 127/250 [04:20<04:10,  2.03s/it]\n",
      "Loss: 0.01940:  51%|█████     | 127/250 [04:20<04:10,  2.03s/it]\n",
      "Loss: 0.02208:  51%|█████     | 127/250 [04:21<04:10,  2.03s/it]\n",
      "Loss: 0.01759:  51%|█████     | 127/250 [04:21<04:10,  2.03s/it]\n",
      "Loss: 0.01906:  51%|█████     | 127/250 [04:21<04:10,  2.03s/it]\n",
      "Loss: 0.01666:  51%|█████     | 127/250 [04:21<04:10,  2.03s/it]\n",
      "Loss: 0.01666:  51%|█████     | 128/250 [04:21<04:09,  2.05s/it]\n",
      "Loss: 0.01916:  51%|█████     | 128/250 [04:21<04:09,  2.05s/it]\n",
      "Loss: 0.01858:  51%|█████     | 128/250 [04:21<04:09,  2.05s/it]\n",
      "Loss: 0.02123:  51%|█████     | 128/250 [04:21<04:09,  2.05s/it]\n",
      "Loss: 0.01953:  51%|█████     | 128/250 [04:21<04:09,  2.05s/it]\n",
      "Loss: 0.01891:  51%|█████     | 128/250 [04:21<04:09,  2.05s/it]\n",
      "Loss: 0.01974:  51%|█████     | 128/250 [04:22<04:09,  2.05s/it]\n",
      "Loss: 0.01819:  51%|█████     | 128/250 [04:22<04:09,  2.05s/it]\n",
      "Loss: 0.01926:  51%|█████     | 128/250 [04:22<04:09,  2.05s/it]\n",
      "Loss: 0.02008:  51%|█████     | 128/250 [04:22<04:09,  2.05s/it]\n",
      "Loss: 0.01902:  51%|█████     | 128/250 [04:22<04:09,  2.05s/it]\n",
      "Loss: 0.01879:  51%|█████     | 128/250 [04:22<04:09,  2.05s/it]\n",
      "Loss: 0.01874:  51%|█████     | 128/250 [04:22<04:09,  2.05s/it]\n",
      "Loss: 0.01951:  51%|█████     | 128/250 [04:22<04:09,  2.05s/it]\n",
      "Loss: 0.01748:  51%|█████     | 128/250 [04:22<04:09,  2.05s/it]\n",
      "Loss: 0.01994:  51%|█████     | 128/250 [04:23<04:09,  2.05s/it]\n",
      "Loss: 0.01996:  51%|█████     | 128/250 [04:23<04:09,  2.05s/it]\n",
      "Loss: 0.02168:  51%|█████     | 128/250 [04:23<04:09,  2.05s/it]\n",
      "Loss: 0.01791:  51%|█████     | 128/250 [04:23<04:09,  2.05s/it]\n",
      "Loss: 0.01647:  51%|█████     | 128/250 [04:23<04:09,  2.05s/it]\n",
      "Loss: 0.01647:  52%|█████▏    | 129/250 [04:23<04:07,  2.05s/it]\n",
      "Loss: 0.01919:  52%|█████▏    | 129/250 [04:23<04:07,  2.05s/it]\n",
      "Loss: 0.01895:  52%|█████▏    | 129/250 [04:23<04:07,  2.05s/it]\n",
      "Loss: 0.02126:  52%|█████▏    | 129/250 [04:23<04:07,  2.05s/it]\n",
      "Loss: 0.01969:  52%|█████▏    | 129/250 [04:23<04:07,  2.05s/it]\n",
      "Loss: 0.01840:  52%|█████▏    | 129/250 [04:24<04:07,  2.05s/it]\n",
      "Loss: 0.01686:  52%|█████▏    | 129/250 [04:24<04:07,  2.05s/it]\n",
      "Loss: 0.01849:  52%|█████▏    | 129/250 [04:24<04:07,  2.05s/it]\n",
      "Loss: 0.02030:  52%|█████▏    | 129/250 [04:24<04:07,  2.05s/it]\n",
      "Loss: 0.01944:  52%|█████▏    | 129/250 [04:24<04:07,  2.05s/it]\n",
      "Loss: 0.02046:  52%|█████▏    | 129/250 [04:24<04:07,  2.05s/it]\n",
      "Loss: 0.02048:  52%|█████▏    | 129/250 [04:24<04:07,  2.05s/it]\n",
      "Loss: 0.02017:  52%|█████▏    | 129/250 [04:24<04:07,  2.05s/it]\n",
      "Loss: 0.01797:  52%|█████▏    | 129/250 [04:24<04:07,  2.05s/it]\n",
      "Loss: 0.01860:  52%|█████▏    | 129/250 [04:25<04:07,  2.05s/it]\n",
      "Loss: 0.01737:  52%|█████▏    | 129/250 [04:25<04:07,  2.05s/it]\n",
      "Loss: 0.01750:  52%|█████▏    | 129/250 [04:25<04:07,  2.05s/it]\n",
      "Loss: 0.02057:  52%|█████▏    | 129/250 [04:25<04:07,  2.05s/it]\n",
      "Loss: 0.01665:  52%|█████▏    | 129/250 [04:25<04:07,  2.05s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.93it/s]\u001b[A\n",
      "Loss: 0.01665:  52%|█████▏    | 130/250 [04:25<04:04,  2.04s/it]\n",
      "Loss: 0.01931:  52%|█████▏    | 130/250 [04:25<04:04,  2.04s/it]\n",
      "Loss: 0.01824:  52%|█████▏    | 130/250 [04:25<04:04,  2.04s/it]\n",
      "Loss: 0.02056:  52%|█████▏    | 130/250 [04:25<04:04,  2.04s/it]\n",
      "Loss: 0.01755:  52%|█████▏    | 130/250 [04:25<04:04,  2.04s/it]\n",
      "Loss: 0.01904:  52%|█████▏    | 130/250 [04:26<04:04,  2.04s/it]\n",
      "Loss: 0.01634:  52%|█████▏    | 130/250 [04:26<04:04,  2.04s/it]\n",
      "Loss: 0.01865:  52%|█████▏    | 130/250 [04:26<04:04,  2.04s/it]\n",
      "Loss: 0.02028:  52%|█████▏    | 130/250 [04:26<04:04,  2.04s/it]\n",
      "Loss: 0.02058:  52%|█████▏    | 130/250 [04:26<04:04,  2.04s/it]\n",
      "Loss: 0.01878:  52%|█████▏    | 130/250 [04:26<04:04,  2.04s/it]\n",
      "Loss: 0.01836:  52%|█████▏    | 130/250 [04:26<04:04,  2.04s/it]\n",
      "Loss: 0.01858:  52%|█████▏    | 130/250 [04:26<04:04,  2.04s/it]\n",
      "Loss: 0.01942:  52%|█████▏    | 130/250 [04:26<04:04,  2.04s/it]\n",
      "Loss: 0.01716:  52%|█████▏    | 130/250 [04:27<04:04,  2.04s/it]\n",
      "Loss: 0.01950:  52%|█████▏    | 130/250 [04:27<04:04,  2.04s/it]\n",
      "Loss: 0.02049:  52%|█████▏    | 130/250 [04:27<04:04,  2.04s/it]\n",
      "Loss: 0.02106:  52%|█████▏    | 130/250 [04:27<04:04,  2.04s/it]\n",
      "Loss: 0.01825:  52%|█████▏    | 130/250 [04:27<04:04,  2.04s/it]\n",
      "Loss: 0.01678:  52%|█████▏    | 130/250 [04:27<04:04,  2.04s/it]\n",
      "Loss: 0.01678:  52%|█████▏    | 131/250 [04:27<04:02,  2.04s/it]\n",
      "Loss: 0.01932:  52%|█████▏    | 131/250 [04:27<04:02,  2.04s/it]\n",
      "Loss: 0.01727:  52%|█████▏    | 131/250 [04:27<04:02,  2.04s/it]\n",
      "Loss: 0.02145:  52%|█████▏    | 131/250 [04:27<04:02,  2.04s/it]\n",
      "Loss: 0.01754:  52%|█████▏    | 131/250 [04:27<04:02,  2.04s/it]\n",
      "Loss: 0.01662:  52%|█████▏    | 131/250 [04:28<04:02,  2.04s/it]\n",
      "Loss: 0.01995:  52%|█████▏    | 131/250 [04:28<04:02,  2.04s/it]\n",
      "Loss: 0.01947:  52%|█████▏    | 131/250 [04:28<04:02,  2.04s/it]\n",
      "Loss: 0.02061:  52%|█████▏    | 131/250 [04:28<04:02,  2.04s/it]\n",
      "Loss: 0.02131:  52%|█████▏    | 131/250 [04:28<04:02,  2.04s/it]\n",
      "Loss: 0.01919:  52%|█████▏    | 131/250 [04:28<04:02,  2.04s/it]\n",
      "Loss: 0.01867:  52%|█████▏    | 131/250 [04:28<04:02,  2.04s/it]\n",
      "Loss: 0.01809:  52%|█████▏    | 131/250 [04:28<04:02,  2.04s/it]\n",
      "Loss: 0.02243:  52%|█████▏    | 131/250 [04:28<04:02,  2.04s/it]\n",
      "Loss: 0.01964:  52%|█████▏    | 131/250 [04:29<04:02,  2.04s/it]\n",
      "Loss: 0.01865:  52%|█████▏    | 131/250 [04:29<04:02,  2.04s/it]\n",
      "Loss: 0.02197:  52%|█████▏    | 131/250 [04:29<04:02,  2.04s/it]\n",
      "Loss: 0.01742:  52%|█████▏    | 131/250 [04:29<04:02,  2.04s/it]\n",
      "Loss: 0.01924:  52%|█████▏    | 131/250 [04:29<04:02,  2.04s/it]\n",
      "Loss: 0.01662:  52%|█████▏    | 131/250 [04:29<04:02,  2.04s/it]\n",
      "Loss: 0.01662:  53%|█████▎    | 132/250 [04:29<04:00,  2.04s/it]\n",
      "Loss: 0.01923:  53%|█████▎    | 132/250 [04:29<04:00,  2.04s/it]\n",
      "Loss: 0.01858:  53%|█████▎    | 132/250 [04:29<04:00,  2.04s/it]\n",
      "Loss: 0.02067:  53%|█████▎    | 132/250 [04:29<04:00,  2.04s/it]\n",
      "Loss: 0.02062:  53%|█████▎    | 132/250 [04:29<04:00,  2.04s/it]\n",
      "Loss: 0.01668:  53%|█████▎    | 132/250 [04:30<04:00,  2.04s/it]\n",
      "Loss: 0.01961:  53%|█████▎    | 132/250 [04:30<04:00,  2.04s/it]\n",
      "Loss: 0.01958:  53%|█████▎    | 132/250 [04:30<04:00,  2.04s/it]\n",
      "Loss: 0.02046:  53%|█████▎    | 132/250 [04:30<04:00,  2.04s/it]\n",
      "Loss: 0.02068:  53%|█████▎    | 132/250 [04:30<04:00,  2.04s/it]\n",
      "Loss: 0.01900:  53%|█████▎    | 132/250 [04:30<04:00,  2.04s/it]\n",
      "Loss: 0.01854:  53%|█████▎    | 132/250 [04:30<04:00,  2.04s/it]\n",
      "Loss: 0.02220:  53%|█████▎    | 132/250 [04:30<04:00,  2.04s/it]\n",
      "Loss: 0.01852:  53%|█████▎    | 132/250 [04:30<04:00,  2.04s/it]\n",
      "Loss: 0.02049:  53%|█████▎    | 132/250 [04:31<04:00,  2.04s/it]\n",
      "Loss: 0.01740:  53%|█████▎    | 132/250 [04:31<04:00,  2.04s/it]\n",
      "Loss: 0.01848:  53%|█████▎    | 132/250 [04:31<04:00,  2.04s/it]\n",
      "Loss: 0.01655:  53%|█████▎    | 132/250 [04:31<04:00,  2.04s/it]\n",
      "Loss: 0.01891:  53%|█████▎    | 132/250 [04:31<04:00,  2.04s/it]\n",
      "Loss: 0.01645:  53%|█████▎    | 132/250 [04:31<04:00,  2.04s/it]\n",
      "Loss: 0.01645:  53%|█████▎    | 133/250 [04:31<03:58,  2.04s/it]\n",
      "Loss: 0.01886:  53%|█████▎    | 133/250 [04:31<03:58,  2.04s/it]\n",
      "Loss: 0.01729:  53%|█████▎    | 133/250 [04:31<03:58,  2.04s/it]\n",
      "Loss: 0.02091:  53%|█████▎    | 133/250 [04:31<03:58,  2.04s/it]\n",
      "Loss: 0.01724:  53%|█████▎    | 133/250 [04:32<03:58,  2.04s/it]\n",
      "Loss: 0.01852:  53%|█████▎    | 133/250 [04:32<03:58,  2.04s/it]\n",
      "Loss: 0.01924:  53%|█████▎    | 133/250 [04:32<03:58,  2.04s/it]\n",
      "Loss: 0.01761:  53%|█████▎    | 133/250 [04:32<03:58,  2.04s/it]\n",
      "Loss: 0.01886:  53%|█████▎    | 133/250 [04:32<03:58,  2.04s/it]\n",
      "Loss: 0.01998:  53%|█████▎    | 133/250 [04:32<03:58,  2.04s/it]\n",
      "Loss: 0.01866:  53%|█████▎    | 133/250 [04:32<03:58,  2.04s/it]\n",
      "Loss: 0.02145:  53%|█████▎    | 133/250 [04:32<03:58,  2.04s/it]\n",
      "Loss: 0.01927:  53%|█████▎    | 133/250 [04:32<03:58,  2.04s/it]\n",
      "Loss: 0.01933:  53%|█████▎    | 133/250 [04:32<03:58,  2.04s/it]\n",
      "Loss: 0.02080:  53%|█████▎    | 133/250 [04:33<03:58,  2.04s/it]\n",
      "Loss: 0.01895:  53%|█████▎    | 133/250 [04:33<03:58,  2.04s/it]\n",
      "Loss: 0.01814:  53%|█████▎    | 133/250 [04:33<03:58,  2.04s/it]\n",
      "Loss: 0.02053:  53%|█████▎    | 133/250 [04:33<03:58,  2.04s/it]\n",
      "Loss: 0.02038:  53%|█████▎    | 133/250 [04:33<03:58,  2.04s/it]\n",
      "Loss: 0.01634:  53%|█████▎    | 133/250 [04:33<03:58,  2.04s/it]\n",
      "Loss: 0.01634:  54%|█████▎    | 134/250 [04:33<03:55,  2.03s/it]\n",
      "Loss: 0.01903:  54%|█████▎    | 134/250 [04:33<03:55,  2.03s/it]\n",
      "Loss: 0.01758:  54%|█████▎    | 134/250 [04:33<03:55,  2.03s/it]\n",
      "Loss: 0.02088:  54%|█████▎    | 134/250 [04:33<03:55,  2.03s/it]\n",
      "Loss: 0.02096:  54%|█████▎    | 134/250 [04:34<03:55,  2.03s/it]\n",
      "Loss: 0.01659:  54%|█████▎    | 134/250 [04:34<03:55,  2.03s/it]\n",
      "Loss: 0.01943:  54%|█████▎    | 134/250 [04:34<03:55,  2.03s/it]\n",
      "Loss: 0.01797:  54%|█████▎    | 134/250 [04:34<03:55,  2.03s/it]\n",
      "Loss: 0.01888:  54%|█████▎    | 134/250 [04:34<03:55,  2.03s/it]\n",
      "Loss: 0.01967:  54%|█████▎    | 134/250 [04:34<03:55,  2.03s/it]\n",
      "Loss: 0.01905:  54%|█████▎    | 134/250 [04:34<03:55,  2.03s/it]\n",
      "Loss: 0.01799:  54%|█████▎    | 134/250 [04:34<03:55,  2.03s/it]\n",
      "Loss: 0.01905:  54%|█████▎    | 134/250 [04:34<03:55,  2.03s/it]\n",
      "Loss: 0.01971:  54%|█████▎    | 134/250 [04:35<03:55,  2.03s/it]\n",
      "Loss: 0.01691:  54%|█████▎    | 134/250 [04:35<03:55,  2.03s/it]\n",
      "Loss: 0.01989:  54%|█████▎    | 134/250 [04:35<03:55,  2.03s/it]\n",
      "Loss: 0.02004:  54%|█████▎    | 134/250 [04:35<03:55,  2.03s/it]\n",
      "Loss: 0.02083:  54%|█████▎    | 134/250 [04:35<03:55,  2.03s/it]\n",
      "Loss: 0.01788:  54%|█████▎    | 134/250 [04:35<03:55,  2.03s/it]\n",
      "Loss: 0.01643:  54%|█████▎    | 134/250 [04:35<03:55,  2.03s/it]\n",
      "Loss: 0.01643:  54%|█████▍    | 135/250 [04:35<03:53,  2.03s/it]\n",
      "Loss: 0.01928:  54%|█████▍    | 135/250 [04:35<03:53,  2.03s/it]\n",
      "Loss: 0.01895:  54%|█████▍    | 135/250 [04:35<03:53,  2.03s/it]\n",
      "Loss: 0.02088:  54%|█████▍    | 135/250 [04:35<03:53,  2.03s/it]\n",
      "Loss: 0.01773:  54%|█████▍    | 135/250 [04:36<03:53,  2.03s/it]\n",
      "Loss: 0.01869:  54%|█████▍    | 135/250 [04:36<03:53,  2.03s/it]\n",
      "Loss: 0.01966:  54%|█████▍    | 135/250 [04:36<03:53,  2.03s/it]\n",
      "Loss: 0.01790:  54%|█████▍    | 135/250 [04:36<03:53,  2.03s/it]\n",
      "Loss: 0.01850:  54%|█████▍    | 135/250 [04:36<03:53,  2.03s/it]\n",
      "Loss: 0.01980:  54%|█████▍    | 135/250 [04:36<03:53,  2.03s/it]\n",
      "Loss: 0.01850:  54%|█████▍    | 135/250 [04:36<03:53,  2.03s/it]\n",
      "Loss: 0.02153:  54%|█████▍    | 135/250 [04:36<03:53,  2.03s/it]\n",
      "Loss: 0.01910:  54%|█████▍    | 135/250 [04:36<03:53,  2.03s/it]\n",
      "Loss: 0.01954:  54%|█████▍    | 135/250 [04:37<03:53,  2.03s/it]\n",
      "Loss: 0.02075:  54%|█████▍    | 135/250 [04:37<03:53,  2.03s/it]\n",
      "Loss: 0.01900:  54%|█████▍    | 135/250 [04:37<03:53,  2.03s/it]\n",
      "Loss: 0.01718:  54%|█████▍    | 135/250 [04:37<03:53,  2.03s/it]\n",
      "Loss: 0.02011:  54%|█████▍    | 135/250 [04:37<03:53,  2.03s/it]\n",
      "Loss: 0.02022:  54%|█████▍    | 135/250 [04:37<03:53,  2.03s/it]\n",
      "Loss: 0.01646:  54%|█████▍    | 135/250 [04:37<03:53,  2.03s/it]\n",
      "Loss: 0.01646:  54%|█████▍    | 136/250 [04:37<03:51,  2.03s/it]\n",
      "Loss: 0.01896:  54%|█████▍    | 136/250 [04:37<03:51,  2.03s/it]\n",
      "Loss: 0.01930:  54%|█████▍    | 136/250 [04:37<03:51,  2.03s/it]\n",
      "Loss: 0.02124:  54%|█████▍    | 136/250 [04:37<03:51,  2.03s/it]\n",
      "Loss: 0.01736:  54%|█████▍    | 136/250 [04:38<03:51,  2.03s/it]\n",
      "Loss: 0.01641:  54%|█████▍    | 136/250 [04:38<03:51,  2.03s/it]\n",
      "Loss: 0.01947:  54%|█████▍    | 136/250 [04:38<03:51,  2.03s/it]\n",
      "Loss: 0.02077:  54%|█████▍    | 136/250 [04:38<03:51,  2.03s/it]\n",
      "Loss: 0.01903:  54%|█████▍    | 136/250 [04:38<03:51,  2.03s/it]\n",
      "Loss: 0.01962:  54%|█████▍    | 136/250 [04:38<03:51,  2.03s/it]\n",
      "Loss: 0.01868:  54%|█████▍    | 136/250 [04:38<03:51,  2.03s/it]\n",
      "Loss: 0.01857:  54%|█████▍    | 136/250 [04:38<03:51,  2.03s/it]\n",
      "Loss: 0.01735:  54%|█████▍    | 136/250 [04:38<03:51,  2.03s/it]\n",
      "Loss: 0.02187:  54%|█████▍    | 136/250 [04:39<03:51,  2.03s/it]\n",
      "Loss: 0.01950:  54%|█████▍    | 136/250 [04:39<03:51,  2.03s/it]\n",
      "Loss: 0.01824:  54%|█████▍    | 136/250 [04:39<03:51,  2.03s/it]\n",
      "Loss: 0.02189:  54%|█████▍    | 136/250 [04:39<03:51,  2.03s/it]\n",
      "Loss: 0.01855:  54%|█████▍    | 136/250 [04:39<03:51,  2.03s/it]\n",
      "Loss: 0.01879:  54%|█████▍    | 136/250 [04:39<03:51,  2.03s/it]\n",
      "Loss: 0.01632:  54%|█████▍    | 136/250 [04:39<03:51,  2.03s/it]\n",
      "Loss: 0.01632:  55%|█████▍    | 137/250 [04:39<03:50,  2.04s/it]\n",
      "Loss: 0.01877:  55%|█████▍    | 137/250 [04:39<03:50,  2.04s/it]\n",
      "Loss: 0.01890:  55%|█████▍    | 137/250 [04:39<03:50,  2.04s/it]\n",
      "Loss: 0.02071:  55%|█████▍    | 137/250 [04:40<03:50,  2.04s/it]\n",
      "Loss: 0.01748:  55%|█████▍    | 137/250 [04:40<03:50,  2.04s/it]\n",
      "Loss: 0.01675:  55%|█████▍    | 137/250 [04:40<03:50,  2.04s/it]\n",
      "Loss: 0.01970:  55%|█████▍    | 137/250 [04:40<03:50,  2.04s/it]\n",
      "Loss: 0.02080:  55%|█████▍    | 137/250 [04:40<03:50,  2.04s/it]\n",
      "Loss: 0.02000:  55%|█████▍    | 137/250 [04:40<03:50,  2.04s/it]\n",
      "Loss: 0.01947:  55%|█████▍    | 137/250 [04:40<03:50,  2.04s/it]\n",
      "Loss: 0.01859:  55%|█████▍    | 137/250 [04:40<03:50,  2.04s/it]\n",
      "Loss: 0.01833:  55%|█████▍    | 137/250 [04:40<03:50,  2.04s/it]\n",
      "Loss: 0.02110:  55%|█████▍    | 137/250 [04:41<03:50,  2.04s/it]\n",
      "Loss: 0.01780:  55%|█████▍    | 137/250 [04:41<03:50,  2.04s/it]\n",
      "Loss: 0.02014:  55%|█████▍    | 137/250 [04:41<03:50,  2.04s/it]\n",
      "Loss: 0.01740:  55%|█████▍    | 137/250 [04:41<03:50,  2.04s/it]\n",
      "Loss: 0.01817:  55%|█████▍    | 137/250 [04:41<03:50,  2.04s/it]\n",
      "Loss: 0.01631:  55%|█████▍    | 137/250 [04:41<03:50,  2.04s/it]\n",
      "Loss: 0.01915:  55%|█████▍    | 137/250 [04:41<03:50,  2.04s/it]\n",
      "Loss: 0.01634:  55%|█████▍    | 137/250 [04:41<03:50,  2.04s/it]\n",
      "Loss: 0.01634:  55%|█████▌    | 138/250 [04:41<03:48,  2.04s/it]\n",
      "Loss: 0.01912:  55%|█████▌    | 138/250 [04:41<03:48,  2.04s/it]\n",
      "Loss: 0.01858:  55%|█████▌    | 138/250 [04:41<03:48,  2.04s/it]\n",
      "Loss: 0.02076:  55%|█████▌    | 138/250 [04:42<03:48,  2.04s/it]\n",
      "Loss: 0.01723:  55%|█████▌    | 138/250 [04:42<03:48,  2.04s/it]\n",
      "Loss: 0.01873:  55%|█████▌    | 138/250 [04:42<03:48,  2.04s/it]\n",
      "Loss: 0.01611:  55%|█████▌    | 138/250 [04:42<03:48,  2.04s/it]\n",
      "Loss: 0.01849:  55%|█████▌    | 138/250 [04:42<03:48,  2.04s/it]\n",
      "Loss: 0.02023:  55%|█████▌    | 138/250 [04:42<03:48,  2.04s/it]\n",
      "Loss: 0.01972:  55%|█████▌    | 138/250 [04:42<03:48,  2.04s/it]\n",
      "Loss: 0.01878:  55%|█████▌    | 138/250 [04:42<03:48,  2.04s/it]\n",
      "Loss: 0.02170:  55%|█████▌    | 138/250 [04:42<03:48,  2.04s/it]\n",
      "Loss: 0.01915:  55%|█████▌    | 138/250 [04:43<03:48,  2.04s/it]\n",
      "Loss: 0.01932:  55%|█████▌    | 138/250 [04:43<03:48,  2.04s/it]\n",
      "Loss: 0.02086:  55%|█████▌    | 138/250 [04:43<03:48,  2.04s/it]\n",
      "Loss: 0.01942:  55%|█████▌    | 138/250 [04:43<03:48,  2.04s/it]\n",
      "Loss: 0.01717:  55%|█████▌    | 138/250 [04:43<03:48,  2.04s/it]\n",
      "Loss: 0.02044:  55%|█████▌    | 138/250 [04:43<03:48,  2.04s/it]\n",
      "Loss: 0.02035:  55%|█████▌    | 138/250 [04:43<03:48,  2.04s/it]\n",
      "Loss: 0.01672:  55%|█████▌    | 138/250 [04:43<03:48,  2.04s/it]\n",
      "Loss: 0.01672:  56%|█████▌    | 139/250 [04:43<03:46,  2.04s/it]\n",
      "Loss: 0.01901:  56%|█████▌    | 139/250 [04:43<03:46,  2.04s/it]\n",
      "Loss: 0.01837:  56%|█████▌    | 139/250 [04:44<03:46,  2.04s/it]\n",
      "Loss: 0.02068:  56%|█████▌    | 139/250 [04:44<03:46,  2.04s/it]\n",
      "Loss: 0.01734:  56%|█████▌    | 139/250 [04:44<03:46,  2.04s/it]\n",
      "Loss: 0.01851:  56%|█████▌    | 139/250 [04:44<03:46,  2.04s/it]\n",
      "Loss: 0.01626:  56%|█████▌    | 139/250 [04:44<03:46,  2.04s/it]\n",
      "Loss: 0.01803:  56%|█████▌    | 139/250 [04:44<03:46,  2.04s/it]\n",
      "Loss: 0.01874:  56%|█████▌    | 139/250 [04:44<03:46,  2.04s/it]\n",
      "Loss: 0.01986:  56%|█████▌    | 139/250 [04:44<03:46,  2.04s/it]\n",
      "Loss: 0.01878:  56%|█████▌    | 139/250 [04:44<03:46,  2.04s/it]\n",
      "Loss: 0.02172:  56%|█████▌    | 139/250 [04:44<03:46,  2.04s/it]\n",
      "Loss: 0.01891:  56%|█████▌    | 139/250 [04:45<03:46,  2.04s/it]\n",
      "Loss: 0.01957:  56%|█████▌    | 139/250 [04:45<03:46,  2.04s/it]\n",
      "Loss: 0.02063:  56%|█████▌    | 139/250 [04:45<03:46,  2.04s/it]\n",
      "Loss: 0.01910:  56%|█████▌    | 139/250 [04:45<03:46,  2.04s/it]\n",
      "Loss: 0.01734:  56%|█████▌    | 139/250 [04:45<03:46,  2.04s/it]\n",
      "Loss: 0.02111:  56%|█████▌    | 139/250 [04:45<03:46,  2.04s/it]\n",
      "Loss: 0.02022:  56%|█████▌    | 139/250 [04:45<03:46,  2.04s/it]\n",
      "Loss: 0.01673:  56%|█████▌    | 139/250 [04:45<03:46,  2.04s/it]\n",
      "Loss: 0.01673:  56%|█████▌    | 140/250 [04:45<03:43,  2.03s/it]\n",
      "Loss: 0.01898:  56%|█████▌    | 140/250 [04:45<03:43,  2.03s/it]\n",
      "Loss: 0.01863:  56%|█████▌    | 140/250 [04:46<03:43,  2.03s/it]\n",
      "Loss: 0.02065:  56%|█████▌    | 140/250 [04:46<03:43,  2.03s/it]\n",
      "Loss: 0.01939:  56%|█████▌    | 140/250 [04:46<03:43,  2.03s/it]\n",
      "Loss: 0.01833:  56%|█████▌    | 140/250 [04:46<03:43,  2.03s/it]\n",
      "Loss: 0.01849:  56%|█████▌    | 140/250 [04:46<03:43,  2.03s/it]\n",
      "Loss: 0.01886:  56%|█████▌    | 140/250 [04:46<03:43,  2.03s/it]\n",
      "Loss: 0.01965:  56%|█████▌    | 140/250 [04:46<03:43,  2.03s/it]\n",
      "Loss: 0.01717:  56%|█████▌    | 140/250 [04:46<03:43,  2.03s/it]\n",
      "Loss: 0.02010:  56%|█████▌    | 140/250 [04:46<03:43,  2.03s/it]\n",
      "Loss: 0.02105:  56%|█████▌    | 140/250 [04:47<03:43,  2.03s/it]\n",
      "Loss: 0.01848:  56%|█████▌    | 140/250 [04:47<03:43,  2.03s/it]\n",
      "Loss: 0.01887:  56%|█████▌    | 140/250 [04:47<03:43,  2.03s/it]\n",
      "Loss: 0.01863:  56%|█████▌    | 140/250 [04:47<03:43,  2.03s/it]\n",
      "Loss: 0.01831:  56%|█████▌    | 140/250 [04:47<03:43,  2.03s/it]\n",
      "Loss: 0.02058:  56%|█████▌    | 140/250 [04:47<03:43,  2.03s/it]\n",
      "Loss: 0.01886:  56%|█████▌    | 140/250 [04:47<03:43,  2.03s/it]\n",
      "Loss: 0.01696:  56%|█████▌    | 140/250 [04:47<03:43,  2.03s/it]\n",
      "Loss: 0.01696:  56%|█████▋    | 141/250 [04:47<03:40,  2.02s/it]\n",
      "Loss: 0.01931:  56%|█████▋    | 141/250 [04:47<03:40,  2.02s/it]\n",
      "Loss: 0.01868:  56%|█████▋    | 141/250 [04:48<03:40,  2.02s/it]\n",
      "Loss: 0.02100:  56%|█████▋    | 141/250 [04:48<03:40,  2.02s/it]\n",
      "Loss: 0.01753:  56%|█████▋    | 141/250 [04:48<03:40,  2.02s/it]\n",
      "Loss: 0.01631:  56%|█████▋    | 141/250 [04:48<03:40,  2.02s/it]\n",
      "Loss: 0.01930:  56%|█████▋    | 141/250 [04:48<03:40,  2.02s/it]\n",
      "Loss: 0.01790:  56%|█████▋    | 141/250 [04:48<03:40,  2.02s/it]\n",
      "Loss: 0.01865:  56%|█████▋    | 141/250 [04:48<03:40,  2.02s/it]\n",
      "Loss: 0.01997:  56%|█████▋    | 141/250 [04:48<03:40,  2.02s/it]\n",
      "Loss: 0.01899:  56%|█████▋    | 141/250 [04:48<03:40,  2.02s/it]\n",
      "Loss: 0.01801:  56%|█████▋    | 141/250 [04:48<03:40,  2.02s/it]\n",
      "Loss: 0.01876:  56%|█████▋    | 141/250 [04:49<03:40,  2.02s/it]\n",
      "Loss: 0.01919:  56%|█████▋    | 141/250 [04:49<03:40,  2.02s/it]\n",
      "Loss: 0.01749:  56%|█████▋    | 141/250 [04:49<03:40,  2.02s/it]\n",
      "Loss: 0.01977:  56%|█████▋    | 141/250 [04:49<03:40,  2.02s/it]\n",
      "Loss: 0.02017:  56%|█████▋    | 141/250 [04:49<03:40,  2.02s/it]\n",
      "Loss: 0.02109:  56%|█████▋    | 141/250 [04:49<03:40,  2.02s/it]\n",
      "Loss: 0.01755:  56%|█████▋    | 141/250 [04:49<03:40,  2.02s/it]\n",
      "Loss: 0.01652:  56%|█████▋    | 141/250 [04:49<03:40,  2.02s/it]\n",
      "Loss: 0.01652:  57%|█████▋    | 142/250 [04:49<03:38,  2.02s/it]\n",
      "Loss: 0.01907:  57%|█████▋    | 142/250 [04:49<03:38,  2.02s/it]\n",
      "Loss: 0.01897:  57%|█████▋    | 142/250 [04:50<03:38,  2.02s/it]\n",
      "Loss: 0.02071:  57%|█████▋    | 142/250 [04:50<03:38,  2.02s/it]\n",
      "Loss: 0.01741:  57%|█████▋    | 142/250 [04:50<03:38,  2.02s/it]\n",
      "Loss: 0.01854:  57%|█████▋    | 142/250 [04:50<03:38,  2.02s/it]\n",
      "Loss: 0.01621:  57%|█████▋    | 142/250 [04:50<03:38,  2.02s/it]\n",
      "Loss: 0.01875:  57%|█████▋    | 142/250 [04:50<03:38,  2.02s/it]\n",
      "Loss: 0.02041:  57%|█████▋    | 142/250 [04:50<03:38,  2.02s/it]\n",
      "Loss: 0.01944:  57%|█████▋    | 142/250 [04:50<03:38,  2.02s/it]\n",
      "Loss: 0.02005:  57%|█████▋    | 142/250 [04:50<03:38,  2.02s/it]\n",
      "Loss: 0.01995:  57%|█████▋    | 142/250 [04:51<03:38,  2.02s/it]\n",
      "Loss: 0.02047:  57%|█████▋    | 142/250 [04:51<03:38,  2.02s/it]\n",
      "Loss: 0.01814:  57%|█████▋    | 142/250 [04:51<03:38,  2.02s/it]\n",
      "Loss: 0.01830:  57%|█████▋    | 142/250 [04:51<03:38,  2.02s/it]\n",
      "Loss: 0.01713:  57%|█████▋    | 142/250 [04:51<03:38,  2.02s/it]\n",
      "Loss: 0.01723:  57%|█████▋    | 142/250 [04:51<03:38,  2.02s/it]\n",
      "Loss: 0.01902:  57%|█████▋    | 142/250 [04:51<03:38,  2.02s/it]\n",
      "Loss: 0.01675:  57%|█████▋    | 142/250 [04:51<03:38,  2.02s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 62.97it/s]\u001b[A\n",
      "Loss: 0.01675:  57%|█████▋    | 143/250 [04:51<03:36,  2.02s/it]\n",
      "Loss: 0.01916:  57%|█████▋    | 143/250 [04:51<03:36,  2.02s/it]\n",
      "Loss: 0.01864:  57%|█████▋    | 143/250 [04:52<03:36,  2.02s/it]\n",
      "Loss: 0.02071:  57%|█████▋    | 143/250 [04:52<03:36,  2.02s/it]\n",
      "Loss: 0.01900:  57%|█████▋    | 143/250 [04:52<03:36,  2.02s/it]\n",
      "Loss: 0.01847:  57%|█████▋    | 143/250 [04:52<03:36,  2.02s/it]\n",
      "Loss: 0.01811:  57%|█████▋    | 143/250 [04:52<03:36,  2.02s/it]\n",
      "Loss: 0.01901:  57%|█████▋    | 143/250 [04:52<03:36,  2.02s/it]\n",
      "Loss: 0.01906:  57%|█████▋    | 143/250 [04:52<03:36,  2.02s/it]\n",
      "Loss: 0.01710:  57%|█████▋    | 143/250 [04:52<03:36,  2.02s/it]\n",
      "Loss: 0.01999:  57%|█████▋    | 143/250 [04:52<03:36,  2.02s/it]\n",
      "Loss: 0.02118:  57%|█████▋    | 143/250 [04:53<03:36,  2.02s/it]\n",
      "Loss: 0.01886:  57%|█████▋    | 143/250 [04:53<03:36,  2.02s/it]\n",
      "Loss: 0.01878:  57%|█████▋    | 143/250 [04:53<03:36,  2.02s/it]\n",
      "Loss: 0.02013:  57%|█████▋    | 143/250 [04:53<03:36,  2.02s/it]\n",
      "Loss: 0.01810:  57%|█████▋    | 143/250 [04:53<03:36,  2.02s/it]\n",
      "Loss: 0.02010:  57%|█████▋    | 143/250 [04:53<03:36,  2.02s/it]\n",
      "Loss: 0.01857:  57%|█████▋    | 143/250 [04:53<03:36,  2.02s/it]\n",
      "Loss: 0.01632:  57%|█████▋    | 143/250 [04:53<03:36,  2.02s/it]\n",
      "Loss: 0.01632:  58%|█████▊    | 144/250 [04:53<03:33,  2.01s/it]\n",
      "Loss: 0.01886:  58%|█████▊    | 144/250 [04:53<03:33,  2.01s/it]\n",
      "Loss: 0.01851:  58%|█████▊    | 144/250 [04:54<03:33,  2.01s/it]\n",
      "Loss: 0.02141:  58%|█████▊    | 144/250 [04:54<03:33,  2.01s/it]\n",
      "Loss: 0.01909:  58%|█████▊    | 144/250 [04:54<03:33,  2.01s/it]\n",
      "Loss: 0.01835:  58%|█████▊    | 144/250 [04:54<03:33,  2.01s/it]\n",
      "Loss: 0.01831:  58%|█████▊    | 144/250 [04:54<03:33,  2.01s/it]\n",
      "Loss: 0.01905:  58%|█████▊    | 144/250 [04:54<03:33,  2.01s/it]\n",
      "Loss: 0.01891:  58%|█████▊    | 144/250 [04:54<03:33,  2.01s/it]\n",
      "Loss: 0.01758:  58%|█████▊    | 144/250 [04:54<03:33,  2.01s/it]\n",
      "Loss: 0.01991:  58%|█████▊    | 144/250 [04:54<03:33,  2.01s/it]\n",
      "Loss: 0.01950:  58%|█████▊    | 144/250 [04:55<03:33,  2.01s/it]\n",
      "Loss: 0.02037:  58%|█████▊    | 144/250 [04:55<03:33,  2.01s/it]\n",
      "Loss: 0.01791:  58%|█████▊    | 144/250 [04:55<03:33,  2.01s/it]\n",
      "Loss: 0.01818:  58%|█████▊    | 144/250 [04:55<03:33,  2.01s/it]\n",
      "Loss: 0.01714:  58%|█████▊    | 144/250 [04:55<03:33,  2.01s/it]\n",
      "Loss: 0.01722:  58%|█████▊    | 144/250 [04:55<03:33,  2.01s/it]\n",
      "Loss: 0.01899:  58%|█████▊    | 144/250 [04:55<03:33,  2.01s/it]\n",
      "Loss: 0.01651:  58%|█████▊    | 144/250 [04:55<03:33,  2.01s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 62.23it/s]\u001b[A\n",
      "Loss: 0.01651:  58%|█████▊    | 145/250 [04:55<03:31,  2.01s/it]\n",
      "Loss: 0.01913:  58%|█████▊    | 145/250 [04:55<03:31,  2.01s/it]\n",
      "Loss: 0.01862:  58%|█████▊    | 145/250 [04:56<03:31,  2.01s/it]\n",
      "Loss: 0.02107:  58%|█████▊    | 145/250 [04:56<03:31,  2.01s/it]\n",
      "Loss: 0.01779:  58%|█████▊    | 145/250 [04:56<03:31,  2.01s/it]\n",
      "Loss: 0.01855:  58%|█████▊    | 145/250 [04:56<03:31,  2.01s/it]\n",
      "Loss: 0.01618:  58%|█████▊    | 145/250 [04:56<03:31,  2.01s/it]\n",
      "Loss: 0.01856:  58%|█████▊    | 145/250 [04:56<03:31,  2.01s/it]\n",
      "Loss: 0.02029:  58%|█████▊    | 145/250 [04:56<03:31,  2.01s/it]\n",
      "Loss: 0.01913:  58%|█████▊    | 145/250 [04:56<03:31,  2.01s/it]\n",
      "Loss: 0.01990:  58%|█████▊    | 145/250 [04:56<03:31,  2.01s/it]\n",
      "Loss: 0.01959:  58%|█████▊    | 145/250 [04:57<03:31,  2.01s/it]\n",
      "Loss: 0.02055:  58%|█████▊    | 145/250 [04:57<03:31,  2.01s/it]\n",
      "Loss: 0.01792:  58%|█████▊    | 145/250 [04:57<03:31,  2.01s/it]\n",
      "Loss: 0.01847:  58%|█████▊    | 145/250 [04:57<03:31,  2.01s/it]\n",
      "Loss: 0.01686:  58%|█████▊    | 145/250 [04:57<03:31,  2.01s/it]\n",
      "Loss: 0.01724:  58%|█████▊    | 145/250 [04:57<03:31,  2.01s/it]\n",
      "Loss: 0.01863:  58%|█████▊    | 145/250 [04:57<03:31,  2.01s/it]\n",
      "Loss: 0.01643:  58%|█████▊    | 145/250 [04:57<03:31,  2.01s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 60.99it/s]\u001b[A\n",
      "Loss: 0.01643:  58%|█████▊    | 146/250 [04:57<03:29,  2.02s/it]\n",
      "Loss: 0.01889:  58%|█████▊    | 146/250 [04:57<03:29,  2.02s/it]\n",
      "Loss: 0.01840:  58%|█████▊    | 146/250 [04:58<03:29,  2.02s/it]\n",
      "Loss: 0.02094:  58%|█████▊    | 146/250 [04:58<03:29,  2.02s/it]\n",
      "Loss: 0.01764:  58%|█████▊    | 146/250 [04:58<03:29,  2.02s/it]\n",
      "Loss: 0.01874:  58%|█████▊    | 146/250 [04:58<03:29,  2.02s/it]\n",
      "Loss: 0.01604:  58%|█████▊    | 146/250 [04:58<03:29,  2.02s/it]\n",
      "Loss: 0.01831:  58%|█████▊    | 146/250 [04:58<03:29,  2.02s/it]\n",
      "Loss: 0.01847:  58%|█████▊    | 146/250 [04:58<03:29,  2.02s/it]\n",
      "Loss: 0.01964:  58%|█████▊    | 146/250 [04:58<03:29,  2.02s/it]\n",
      "Loss: 0.01907:  58%|█████▊    | 146/250 [04:58<03:29,  2.02s/it]\n",
      "Loss: 0.01826:  58%|█████▊    | 146/250 [04:59<03:29,  2.02s/it]\n",
      "Loss: 0.01831:  58%|█████▊    | 146/250 [04:59<03:29,  2.02s/it]\n",
      "Loss: 0.01966:  58%|█████▊    | 146/250 [04:59<03:29,  2.02s/it]\n",
      "Loss: 0.01675:  58%|█████▊    | 146/250 [04:59<03:29,  2.02s/it]\n",
      "Loss: 0.01980:  58%|█████▊    | 146/250 [04:59<03:29,  2.02s/it]\n",
      "Loss: 0.02028:  58%|█████▊    | 146/250 [04:59<03:29,  2.02s/it]\n",
      "Loss: 0.02071:  58%|█████▊    | 146/250 [04:59<03:29,  2.02s/it]\n",
      "Loss: 0.01752:  58%|█████▊    | 146/250 [04:59<03:29,  2.02s/it]\n",
      "Loss: 0.01666:  58%|█████▊    | 146/250 [04:59<03:29,  2.02s/it]\n",
      "Loss: 0.01666:  59%|█████▉    | 147/250 [04:59<03:28,  2.02s/it]\n",
      "Loss: 0.01876:  59%|█████▉    | 147/250 [05:00<03:28,  2.02s/it]\n",
      "Loss: 0.01849:  59%|█████▉    | 147/250 [05:00<03:28,  2.02s/it]\n",
      "Loss: 0.02028:  59%|█████▉    | 147/250 [05:00<03:28,  2.02s/it]\n",
      "Loss: 0.01902:  59%|█████▉    | 147/250 [05:00<03:28,  2.02s/it]\n",
      "Loss: 0.01837:  59%|█████▉    | 147/250 [05:00<03:28,  2.02s/it]\n",
      "Loss: 0.01625:  59%|█████▉    | 147/250 [05:00<03:28,  2.02s/it]\n",
      "Loss: 0.01807:  59%|█████▉    | 147/250 [05:00<03:28,  2.02s/it]\n",
      "Loss: 0.01834:  59%|█████▉    | 147/250 [05:00<03:28,  2.02s/it]\n",
      "Loss: 0.01967:  59%|█████▉    | 147/250 [05:00<03:28,  2.02s/it]\n",
      "Loss: 0.01837:  59%|█████▉    | 147/250 [05:00<03:28,  2.02s/it]\n",
      "Loss: 0.02125:  59%|█████▉    | 147/250 [05:01<03:28,  2.02s/it]\n",
      "Loss: 0.01904:  59%|█████▉    | 147/250 [05:01<03:28,  2.02s/it]\n",
      "Loss: 0.01920:  59%|█████▉    | 147/250 [05:01<03:28,  2.02s/it]\n",
      "Loss: 0.02045:  59%|█████▉    | 147/250 [05:01<03:28,  2.02s/it]\n",
      "Loss: 0.01893:  59%|█████▉    | 147/250 [05:01<03:28,  2.02s/it]\n",
      "Loss: 0.01692:  59%|█████▉    | 147/250 [05:01<03:28,  2.02s/it]\n",
      "Loss: 0.02043:  59%|█████▉    | 147/250 [05:01<03:28,  2.02s/it]\n",
      "Loss: 0.02102:  59%|█████▉    | 147/250 [05:01<03:28,  2.02s/it]\n",
      "Loss: 0.01662:  59%|█████▉    | 147/250 [05:01<03:28,  2.02s/it]\n",
      "Loss: 0.01662:  59%|█████▉    | 148/250 [05:01<03:25,  2.02s/it]\n",
      "Loss: 0.01805:  59%|█████▉    | 148/250 [05:02<03:25,  2.02s/it]\n",
      "Loss: 0.01738:  59%|█████▉    | 148/250 [05:02<03:25,  2.02s/it]\n",
      "Loss: 0.02066:  59%|█████▉    | 148/250 [05:02<03:25,  2.02s/it]\n",
      "Loss: 0.01729:  59%|█████▉    | 148/250 [05:02<03:25,  2.02s/it]\n",
      "Loss: 0.01829:  59%|█████▉    | 148/250 [05:02<03:25,  2.02s/it]\n",
      "Loss: 0.01593:  59%|█████▉    | 148/250 [05:02<03:25,  2.02s/it]\n",
      "Loss: 0.01843:  59%|█████▉    | 148/250 [05:02<03:25,  2.02s/it]\n",
      "Loss: 0.02015:  59%|█████▉    | 148/250 [05:02<03:25,  2.02s/it]\n",
      "Loss: 0.01939:  59%|█████▉    | 148/250 [05:02<03:25,  2.02s/it]\n",
      "Loss: 0.01898:  59%|█████▉    | 148/250 [05:03<03:25,  2.02s/it]\n",
      "Loss: 0.01791:  59%|█████▉    | 148/250 [05:03<03:25,  2.02s/it]\n",
      "Loss: 0.01812:  59%|█████▉    | 148/250 [05:03<03:25,  2.02s/it]\n",
      "Loss: 0.01908:  59%|█████▉    | 148/250 [05:03<03:25,  2.02s/it]\n",
      "Loss: 0.01679:  59%|█████▉    | 148/250 [05:03<03:25,  2.02s/it]\n",
      "Loss: 0.01935:  59%|█████▉    | 148/250 [05:03<03:25,  2.02s/it]\n",
      "Loss: 0.01996:  59%|█████▉    | 148/250 [05:03<03:25,  2.02s/it]\n",
      "Loss: 0.02075:  59%|█████▉    | 148/250 [05:03<03:25,  2.02s/it]\n",
      "Loss: 0.01776:  59%|█████▉    | 148/250 [05:03<03:25,  2.02s/it]\n",
      "Loss: 0.01651:  59%|█████▉    | 148/250 [05:03<03:25,  2.02s/it]\n",
      "Loss: 0.01651:  60%|█████▉    | 149/250 [05:03<03:24,  2.02s/it]\n",
      "Loss: 0.01890:  60%|█████▉    | 149/250 [05:04<03:24,  2.02s/it]\n",
      "Loss: 0.01850:  60%|█████▉    | 149/250 [05:04<03:24,  2.02s/it]\n",
      "Loss: 0.02065:  60%|█████▉    | 149/250 [05:04<03:24,  2.02s/it]\n",
      "Loss: 0.01964:  60%|█████▉    | 149/250 [05:04<03:24,  2.02s/it]\n",
      "Loss: 0.01914:  60%|█████▉    | 149/250 [05:04<03:24,  2.02s/it]\n",
      "Loss: 0.01830:  60%|█████▉    | 149/250 [05:04<03:24,  2.02s/it]\n",
      "Loss: 0.01940:  60%|█████▉    | 149/250 [05:04<03:24,  2.02s/it]\n",
      "Loss: 0.02040:  60%|█████▉    | 149/250 [05:04<03:24,  2.02s/it]\n",
      "Loss: 0.01883:  60%|█████▉    | 149/250 [05:04<03:24,  2.02s/it]\n",
      "Loss: 0.01890:  60%|█████▉    | 149/250 [05:05<03:24,  2.02s/it]\n",
      "Loss: 0.02138:  60%|█████▉    | 149/250 [05:05<03:24,  2.02s/it]\n",
      "Loss: 0.01941:  60%|█████▉    | 149/250 [05:05<03:24,  2.02s/it]\n",
      "Loss: 0.01910:  60%|█████▉    | 149/250 [05:05<03:24,  2.02s/it]\n",
      "Loss: 0.01710:  60%|█████▉    | 149/250 [05:05<03:24,  2.02s/it]\n",
      "Loss: 0.01964:  60%|█████▉    | 149/250 [05:05<03:24,  2.02s/it]\n",
      "Loss: 0.01996:  60%|█████▉    | 149/250 [05:05<03:24,  2.02s/it]\n",
      "Loss: 0.02112:  60%|█████▉    | 149/250 [05:05<03:24,  2.02s/it]\n",
      "Loss: 0.01736:  60%|█████▉    | 149/250 [05:05<03:24,  2.02s/it]\n",
      "Loss: 0.01650:  60%|█████▉    | 149/250 [05:05<03:24,  2.02s/it]\n",
      "Loss: 0.01650:  60%|██████    | 150/250 [05:06<03:23,  2.03s/it]\n",
      "Loss: 0.01882:  60%|██████    | 150/250 [05:06<03:23,  2.03s/it]\n",
      "Loss: 0.01839:  60%|██████    | 150/250 [05:06<03:23,  2.03s/it]\n",
      "Loss: 0.02135:  60%|██████    | 150/250 [05:06<03:23,  2.03s/it]\n",
      "Loss: 0.01906:  60%|██████    | 150/250 [05:06<03:23,  2.03s/it]\n",
      "Loss: 0.01860:  60%|██████    | 150/250 [05:06<03:23,  2.03s/it]\n",
      "Loss: 0.01809:  60%|██████    | 150/250 [05:06<03:23,  2.03s/it]\n",
      "Loss: 0.01807:  60%|██████    | 150/250 [05:06<03:23,  2.03s/it]\n",
      "Loss: 0.01998:  60%|██████    | 150/250 [05:06<03:23,  2.03s/it]\n",
      "Loss: 0.01904:  60%|██████    | 150/250 [05:06<03:23,  2.03s/it]\n",
      "Loss: 0.01966:  60%|██████    | 150/250 [05:07<03:23,  2.03s/it]\n",
      "Loss: 0.01963:  60%|██████    | 150/250 [05:07<03:23,  2.03s/it]\n",
      "Loss: 0.01957:  60%|██████    | 150/250 [05:07<03:23,  2.03s/it]\n",
      "Loss: 0.01753:  60%|██████    | 150/250 [05:07<03:23,  2.03s/it]\n",
      "Loss: 0.01802:  60%|██████    | 150/250 [05:07<03:23,  2.03s/it]\n",
      "Loss: 0.01685:  60%|██████    | 150/250 [05:07<03:23,  2.03s/it]\n",
      "Loss: 0.01754:  60%|██████    | 150/250 [05:07<03:23,  2.03s/it]\n",
      "Loss: 0.01854:  60%|██████    | 150/250 [05:07<03:23,  2.03s/it]\n",
      "Loss: 0.01626:  60%|██████    | 150/250 [05:08<03:23,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.95it/s]\u001b[A\n",
      "Loss: 0.01626:  60%|██████    | 151/250 [05:08<03:20,  2.03s/it]\n",
      "Loss: 0.01880:  60%|██████    | 151/250 [05:08<03:20,  2.03s/it]\n",
      "Loss: 0.01834:  60%|██████    | 151/250 [05:08<03:20,  2.03s/it]\n",
      "Loss: 0.02051:  60%|██████    | 151/250 [05:08<03:20,  2.03s/it]\n",
      "Loss: 0.01744:  60%|██████    | 151/250 [05:08<03:20,  2.03s/it]\n",
      "Loss: 0.01893:  60%|██████    | 151/250 [05:08<03:20,  2.03s/it]\n",
      "Loss: 0.01694:  60%|██████    | 151/250 [05:08<03:20,  2.03s/it]\n",
      "Loss: 0.01841:  60%|██████    | 151/250 [05:08<03:20,  2.03s/it]\n",
      "Loss: 0.02019:  60%|██████    | 151/250 [05:08<03:20,  2.03s/it]\n",
      "Loss: 0.01887:  60%|██████    | 151/250 [05:09<03:20,  2.03s/it]\n",
      "Loss: 0.01831:  60%|██████    | 151/250 [05:09<03:20,  2.03s/it]\n",
      "Loss: 0.02128:  60%|██████    | 151/250 [05:09<03:20,  2.03s/it]\n",
      "Loss: 0.01895:  60%|██████    | 151/250 [05:09<03:20,  2.03s/it]\n",
      "Loss: 0.01988:  60%|██████    | 151/250 [05:09<03:20,  2.03s/it]\n",
      "Loss: 0.02063:  60%|██████    | 151/250 [05:09<03:20,  2.03s/it]\n",
      "Loss: 0.01892:  60%|██████    | 151/250 [05:09<03:20,  2.03s/it]\n",
      "Loss: 0.01679:  60%|██████    | 151/250 [05:09<03:20,  2.03s/it]\n",
      "Loss: 0.01983:  60%|██████    | 151/250 [05:09<03:20,  2.03s/it]\n",
      "Loss: 0.02026:  60%|██████    | 151/250 [05:10<03:20,  2.03s/it]\n",
      "Loss: 0.01682:  60%|██████    | 151/250 [05:10<03:20,  2.03s/it]\n",
      "Loss: 0.01682:  61%|██████    | 152/250 [05:10<03:18,  2.02s/it]\n",
      "Loss: 0.01884:  61%|██████    | 152/250 [05:10<03:18,  2.02s/it]\n",
      "Loss: 0.01854:  61%|██████    | 152/250 [05:10<03:18,  2.02s/it]\n",
      "Loss: 0.02072:  61%|██████    | 152/250 [05:10<03:18,  2.02s/it]\n",
      "Loss: 0.01895:  61%|██████    | 152/250 [05:10<03:18,  2.02s/it]\n",
      "Loss: 0.01810:  61%|██████    | 152/250 [05:10<03:18,  2.02s/it]\n",
      "Loss: 0.01801:  61%|██████    | 152/250 [05:10<03:18,  2.02s/it]\n",
      "Loss: 0.01884:  61%|██████    | 152/250 [05:10<03:18,  2.02s/it]\n",
      "Loss: 0.01930:  61%|██████    | 152/250 [05:10<03:18,  2.02s/it]\n",
      "Loss: 0.01698:  61%|██████    | 152/250 [05:11<03:18,  2.02s/it]\n",
      "Loss: 0.02015:  61%|██████    | 152/250 [05:11<03:18,  2.02s/it]\n",
      "Loss: 0.02142:  61%|██████    | 152/250 [05:11<03:18,  2.02s/it]\n",
      "Loss: 0.01837:  61%|██████    | 152/250 [05:11<03:18,  2.02s/it]\n",
      "Loss: 0.01829:  61%|██████    | 152/250 [05:11<03:18,  2.02s/it]\n",
      "Loss: 0.01913:  61%|██████    | 152/250 [05:11<03:18,  2.02s/it]\n",
      "Loss: 0.01804:  61%|██████    | 152/250 [05:11<03:18,  2.02s/it]\n",
      "Loss: 0.02044:  61%|██████    | 152/250 [05:11<03:18,  2.02s/it]\n",
      "Loss: 0.01862:  61%|██████    | 152/250 [05:11<03:18,  2.02s/it]\n",
      "Loss: 0.01631:  61%|██████    | 152/250 [05:12<03:18,  2.02s/it]\n",
      "Loss: 0.01631:  61%|██████    | 153/250 [05:12<03:15,  2.02s/it]\n",
      "Loss: 0.01877:  61%|██████    | 153/250 [05:12<03:15,  2.02s/it]\n",
      "Loss: 0.01796:  61%|██████    | 153/250 [05:12<03:15,  2.02s/it]\n",
      "Loss: 0.02055:  61%|██████    | 153/250 [05:12<03:15,  2.02s/it]\n",
      "Loss: 0.01741:  61%|██████    | 153/250 [05:12<03:15,  2.02s/it]\n",
      "Loss: 0.01836:  61%|██████    | 153/250 [05:12<03:15,  2.02s/it]\n",
      "Loss: 0.01610:  61%|██████    | 153/250 [05:12<03:15,  2.02s/it]\n",
      "Loss: 0.01838:  61%|██████    | 153/250 [05:12<03:15,  2.02s/it]\n",
      "Loss: 0.01869:  61%|██████    | 153/250 [05:12<03:15,  2.02s/it]\n",
      "Loss: 0.01946:  61%|██████    | 153/250 [05:13<03:15,  2.02s/it]\n",
      "Loss: 0.01884:  61%|██████    | 153/250 [05:13<03:15,  2.02s/it]\n",
      "Loss: 0.01793:  61%|██████    | 153/250 [05:13<03:15,  2.02s/it]\n",
      "Loss: 0.01838:  61%|██████    | 153/250 [05:13<03:15,  2.02s/it]\n",
      "Loss: 0.01944:  61%|██████    | 153/250 [05:13<03:15,  2.02s/it]\n",
      "Loss: 0.01703:  61%|██████    | 153/250 [05:13<03:15,  2.02s/it]\n",
      "Loss: 0.01953:  61%|██████    | 153/250 [05:13<03:15,  2.02s/it]\n",
      "Loss: 0.01974:  61%|██████    | 153/250 [05:13<03:15,  2.02s/it]\n",
      "Loss: 0.02054:  61%|██████    | 153/250 [05:13<03:15,  2.02s/it]\n",
      "Loss: 0.01744:  61%|██████    | 153/250 [05:14<03:15,  2.02s/it]\n",
      "Loss: 0.01623:  61%|██████    | 153/250 [05:14<03:15,  2.02s/it]\n",
      "Loss: 0.01623:  62%|██████▏   | 154/250 [05:14<03:14,  2.03s/it]\n",
      "Loss: 0.01879:  62%|██████▏   | 154/250 [05:14<03:14,  2.03s/it]\n",
      "Loss: 0.01709:  62%|██████▏   | 154/250 [05:14<03:14,  2.03s/it]\n",
      "Loss: 0.02081:  62%|██████▏   | 154/250 [05:14<03:14,  2.03s/it]\n",
      "Loss: 0.01725:  62%|██████▏   | 154/250 [05:14<03:14,  2.03s/it]\n",
      "Loss: 0.01827:  62%|██████▏   | 154/250 [05:14<03:14,  2.03s/it]\n",
      "Loss: 0.01621:  62%|██████▏   | 154/250 [05:14<03:14,  2.03s/it]\n",
      "Loss: 0.01822:  62%|██████▏   | 154/250 [05:14<03:14,  2.03s/it]\n",
      "Loss: 0.02008:  62%|██████▏   | 154/250 [05:14<03:14,  2.03s/it]\n",
      "Loss: 0.01908:  62%|██████▏   | 154/250 [05:15<03:14,  2.03s/it]\n",
      "Loss: 0.01980:  62%|██████▏   | 154/250 [05:15<03:14,  2.03s/it]\n",
      "Loss: 0.01968:  62%|██████▏   | 154/250 [05:15<03:14,  2.03s/it]\n",
      "Loss: 0.02042:  62%|██████▏   | 154/250 [05:15<03:14,  2.03s/it]\n",
      "Loss: 0.01732:  62%|██████▏   | 154/250 [05:15<03:14,  2.03s/it]\n",
      "Loss: 0.01819:  62%|██████▏   | 154/250 [05:15<03:14,  2.03s/it]\n",
      "Loss: 0.01688:  62%|██████▏   | 154/250 [05:15<03:14,  2.03s/it]\n",
      "Loss: 0.01696:  62%|██████▏   | 154/250 [05:15<03:14,  2.03s/it]\n",
      "Loss: 0.01900:  62%|██████▏   | 154/250 [05:15<03:14,  2.03s/it]\n",
      "Loss: 0.01636:  62%|██████▏   | 154/250 [05:16<03:14,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 62.66it/s]\u001b[A\n",
      "Loss: 0.01636:  62%|██████▏   | 155/250 [05:16<03:12,  2.03s/it]\n",
      "Loss: 0.01877:  62%|██████▏   | 155/250 [05:16<03:12,  2.03s/it]\n",
      "Loss: 0.01818:  62%|██████▏   | 155/250 [05:16<03:12,  2.03s/it]\n",
      "Loss: 0.02120:  62%|██████▏   | 155/250 [05:16<03:12,  2.03s/it]\n",
      "Loss: 0.01894:  62%|██████▏   | 155/250 [05:16<03:12,  2.03s/it]\n",
      "Loss: 0.01850:  62%|██████▏   | 155/250 [05:16<03:12,  2.03s/it]\n",
      "Loss: 0.01610:  62%|██████▏   | 155/250 [05:16<03:12,  2.03s/it]\n",
      "Loss: 0.01827:  62%|██████▏   | 155/250 [05:16<03:12,  2.03s/it]\n",
      "Loss: 0.01984:  62%|██████▏   | 155/250 [05:16<03:12,  2.03s/it]\n",
      "Loss: 0.01943:  62%|██████▏   | 155/250 [05:17<03:12,  2.03s/it]\n",
      "Loss: 0.01965:  62%|██████▏   | 155/250 [05:17<03:12,  2.03s/it]\n",
      "Loss: 0.01935:  62%|██████▏   | 155/250 [05:17<03:12,  2.03s/it]\n",
      "Loss: 0.02007:  62%|██████▏   | 155/250 [05:17<03:12,  2.03s/it]\n",
      "Loss: 0.01797:  62%|██████▏   | 155/250 [05:17<03:12,  2.03s/it]\n",
      "Loss: 0.01775:  62%|██████▏   | 155/250 [05:17<03:12,  2.03s/it]\n",
      "Loss: 0.01665:  62%|██████▏   | 155/250 [05:17<03:12,  2.03s/it]\n",
      "Loss: 0.01706:  62%|██████▏   | 155/250 [05:17<03:12,  2.03s/it]\n",
      "Loss: 0.01894:  62%|██████▏   | 155/250 [05:18<03:12,  2.03s/it]\n",
      "Loss: 0.01634:  62%|██████▏   | 155/250 [05:18<03:12,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.37it/s]\u001b[A\n",
      "Loss: 0.01634:  62%|██████▏   | 156/250 [05:18<03:10,  2.03s/it]\n",
      "Loss: 0.01818:  62%|██████▏   | 156/250 [05:18<03:10,  2.03s/it]\n",
      "Loss: 0.01736:  62%|██████▏   | 156/250 [05:18<03:10,  2.03s/it]\n",
      "Loss: 0.02051:  62%|██████▏   | 156/250 [05:18<03:10,  2.03s/it]\n",
      "Loss: 0.01712:  62%|██████▏   | 156/250 [05:18<03:10,  2.03s/it]\n",
      "Loss: 0.01840:  62%|██████▏   | 156/250 [05:18<03:10,  2.03s/it]\n",
      "Loss: 0.01609:  62%|██████▏   | 156/250 [05:18<03:10,  2.03s/it]\n",
      "Loss: 0.01796:  62%|██████▏   | 156/250 [05:18<03:10,  2.03s/it]\n",
      "Loss: 0.02045:  62%|██████▏   | 156/250 [05:19<03:10,  2.03s/it]\n",
      "Loss: 0.01894:  62%|██████▏   | 156/250 [05:19<03:10,  2.03s/it]\n",
      "Loss: 0.01996:  62%|██████▏   | 156/250 [05:19<03:10,  2.03s/it]\n",
      "Loss: 0.01934:  62%|██████▏   | 156/250 [05:19<03:10,  2.03s/it]\n",
      "Loss: 0.02003:  62%|██████▏   | 156/250 [05:19<03:10,  2.03s/it]\n",
      "Loss: 0.01765:  62%|██████▏   | 156/250 [05:19<03:10,  2.03s/it]\n",
      "Loss: 0.01810:  62%|██████▏   | 156/250 [05:19<03:10,  2.03s/it]\n",
      "Loss: 0.01688:  62%|██████▏   | 156/250 [05:19<03:10,  2.03s/it]\n",
      "Loss: 0.01709:  62%|██████▏   | 156/250 [05:19<03:10,  2.03s/it]\n",
      "Loss: 0.01864:  62%|██████▏   | 156/250 [05:20<03:10,  2.03s/it]\n",
      "Loss: 0.01635:  62%|██████▏   | 156/250 [05:20<03:10,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.22it/s]\u001b[A\n",
      "Loss: 0.01635:  63%|██████▎   | 157/250 [05:20<03:08,  2.03s/it]\n",
      "Loss: 0.01905:  63%|██████▎   | 157/250 [05:20<03:08,  2.03s/it]\n",
      "Loss: 0.01829:  63%|██████▎   | 157/250 [05:20<03:08,  2.03s/it]\n",
      "Loss: 0.02037:  63%|██████▎   | 157/250 [05:20<03:08,  2.03s/it]\n",
      "Loss: 0.01724:  63%|██████▎   | 157/250 [05:20<03:08,  2.03s/it]\n",
      "Loss: 0.01844:  63%|██████▎   | 157/250 [05:20<03:08,  2.03s/it]\n",
      "Loss: 0.01630:  63%|██████▎   | 157/250 [05:20<03:08,  2.03s/it]\n",
      "Loss: 0.01781:  63%|██████▎   | 157/250 [05:20<03:08,  2.03s/it]\n",
      "Loss: 0.01845:  63%|██████▎   | 157/250 [05:21<03:08,  2.03s/it]\n",
      "Loss: 0.01930:  63%|██████▎   | 157/250 [05:21<03:08,  2.03s/it]\n",
      "Loss: 0.01843:  63%|██████▎   | 157/250 [05:21<03:08,  2.03s/it]\n",
      "Loss: 0.01796:  63%|██████▎   | 157/250 [05:21<03:08,  2.03s/it]\n",
      "Loss: 0.01795:  63%|██████▎   | 157/250 [05:21<03:08,  2.03s/it]\n",
      "Loss: 0.01879:  63%|██████▎   | 157/250 [05:21<03:08,  2.03s/it]\n",
      "Loss: 0.01681:  63%|██████▎   | 157/250 [05:21<03:08,  2.03s/it]\n",
      "Loss: 0.02011:  63%|██████▎   | 157/250 [05:21<03:08,  2.03s/it]\n",
      "Loss: 0.02013:  63%|██████▎   | 157/250 [05:21<03:08,  2.03s/it]\n",
      "Loss: 0.02086:  63%|██████▎   | 157/250 [05:22<03:08,  2.03s/it]\n",
      "Loss: 0.01746:  63%|██████▎   | 157/250 [05:22<03:08,  2.03s/it]\n",
      "Loss: 0.01659:  63%|██████▎   | 157/250 [05:22<03:08,  2.03s/it]\n",
      "Loss: 0.01659:  63%|██████▎   | 158/250 [05:22<03:06,  2.03s/it]\n",
      "Loss: 0.01974:  63%|██████▎   | 158/250 [05:22<03:06,  2.03s/it]\n",
      "Loss: 0.01854:  63%|██████▎   | 158/250 [05:22<03:06,  2.03s/it]\n",
      "Loss: 0.02008:  63%|██████▎   | 158/250 [05:22<03:06,  2.03s/it]\n",
      "Loss: 0.01941:  63%|██████▎   | 158/250 [05:22<03:06,  2.03s/it]\n",
      "Loss: 0.01775:  63%|██████▎   | 158/250 [05:22<03:06,  2.03s/it]\n",
      "Loss: 0.01797:  63%|██████▎   | 158/250 [05:22<03:06,  2.03s/it]\n",
      "Loss: 0.01890:  63%|██████▎   | 158/250 [05:23<03:06,  2.03s/it]\n",
      "Loss: 0.01884:  63%|██████▎   | 158/250 [05:23<03:06,  2.03s/it]\n",
      "Loss: 0.01883:  63%|██████▎   | 158/250 [05:23<03:06,  2.03s/it]\n",
      "Loss: 0.01853:  63%|██████▎   | 158/250 [05:23<03:06,  2.03s/it]\n",
      "Loss: 0.02063:  63%|██████▎   | 158/250 [05:23<03:06,  2.03s/it]\n",
      "Loss: 0.01926:  63%|██████▎   | 158/250 [05:23<03:06,  2.03s/it]\n",
      "Loss: 0.01929:  63%|██████▎   | 158/250 [05:23<03:06,  2.03s/it]\n",
      "Loss: 0.02068:  63%|██████▎   | 158/250 [05:23<03:06,  2.03s/it]\n",
      "Loss: 0.01834:  63%|██████▎   | 158/250 [05:23<03:06,  2.03s/it]\n",
      "Loss: 0.01750:  63%|██████▎   | 158/250 [05:23<03:06,  2.03s/it]\n",
      "Loss: 0.01985:  63%|██████▎   | 158/250 [05:24<03:06,  2.03s/it]\n",
      "Loss: 0.02007:  63%|██████▎   | 158/250 [05:24<03:06,  2.03s/it]\n",
      "Loss: 0.01611:  63%|██████▎   | 158/250 [05:24<03:06,  2.03s/it]\n",
      "Loss: 0.01611:  64%|██████▎   | 159/250 [05:24<03:04,  2.03s/it]\n",
      "Loss: 0.01898:  64%|██████▎   | 159/250 [05:24<03:04,  2.03s/it]\n",
      "Loss: 0.01835:  64%|██████▎   | 159/250 [05:24<03:04,  2.03s/it]\n",
      "Loss: 0.02013:  64%|██████▎   | 159/250 [05:24<03:04,  2.03s/it]\n",
      "Loss: 0.01734:  64%|██████▎   | 159/250 [05:24<03:04,  2.03s/it]\n",
      "Loss: 0.01826:  64%|██████▎   | 159/250 [05:24<03:04,  2.03s/it]\n",
      "Loss: 0.01906:  64%|██████▎   | 159/250 [05:24<03:04,  2.03s/it]\n",
      "Loss: 0.01731:  64%|██████▎   | 159/250 [05:25<03:04,  2.03s/it]\n",
      "Loss: 0.01862:  64%|██████▎   | 159/250 [05:25<03:04,  2.03s/it]\n",
      "Loss: 0.01931:  64%|██████▎   | 159/250 [05:25<03:04,  2.03s/it]\n",
      "Loss: 0.01860:  64%|██████▎   | 159/250 [05:25<03:04,  2.03s/it]\n",
      "Loss: 0.01802:  64%|██████▎   | 159/250 [05:25<03:04,  2.03s/it]\n",
      "Loss: 0.02037:  64%|██████▎   | 159/250 [05:25<03:04,  2.03s/it]\n",
      "Loss: 0.01784:  64%|██████▎   | 159/250 [05:25<03:04,  2.03s/it]\n",
      "Loss: 0.01930:  64%|██████▎   | 159/250 [05:25<03:04,  2.03s/it]\n",
      "Loss: 0.01711:  64%|██████▎   | 159/250 [05:25<03:04,  2.03s/it]\n",
      "Loss: 0.01805:  64%|██████▎   | 159/250 [05:26<03:04,  2.03s/it]\n",
      "Loss: 0.01558:  64%|██████▎   | 159/250 [05:26<03:04,  2.03s/it]\n",
      "Loss: 0.01863:  64%|██████▎   | 159/250 [05:26<03:04,  2.03s/it]\n",
      "Loss: 0.01627:  64%|██████▎   | 159/250 [05:26<03:04,  2.03s/it]\n",
      "Loss: 0.01627:  64%|██████▍   | 160/250 [05:26<03:03,  2.03s/it]\n",
      "Loss: 0.01845:  64%|██████▍   | 160/250 [05:26<03:03,  2.03s/it]\n",
      "Loss: 0.01876:  64%|██████▍   | 160/250 [05:26<03:03,  2.03s/it]\n",
      "Loss: 0.01998:  64%|██████▍   | 160/250 [05:26<03:03,  2.03s/it]\n",
      "Loss: 0.01691:  64%|██████▍   | 160/250 [05:26<03:03,  2.03s/it]\n",
      "Loss: 0.01596:  64%|██████▍   | 160/250 [05:26<03:03,  2.03s/it]\n",
      "Loss: 0.01879:  64%|██████▍   | 160/250 [05:26<03:03,  2.03s/it]\n",
      "Loss: 0.01695:  64%|██████▍   | 160/250 [05:27<03:03,  2.03s/it]\n",
      "Loss: 0.01866:  64%|██████▍   | 160/250 [05:27<03:03,  2.03s/it]\n",
      "Loss: 0.01921:  64%|██████▍   | 160/250 [05:27<03:03,  2.03s/it]\n",
      "Loss: 0.01905:  64%|██████▍   | 160/250 [05:27<03:03,  2.03s/it]\n",
      "Loss: 0.01800:  64%|██████▍   | 160/250 [05:27<03:03,  2.03s/it]\n",
      "Loss: 0.01868:  64%|██████▍   | 160/250 [05:27<03:03,  2.03s/it]\n",
      "Loss: 0.01890:  64%|██████▍   | 160/250 [05:27<03:03,  2.03s/it]\n",
      "Loss: 0.01655:  64%|██████▍   | 160/250 [05:27<03:03,  2.03s/it]\n",
      "Loss: 0.01899:  64%|██████▍   | 160/250 [05:27<03:03,  2.03s/it]\n",
      "Loss: 0.01948:  64%|██████▍   | 160/250 [05:28<03:03,  2.03s/it]\n",
      "Loss: 0.02059:  64%|██████▍   | 160/250 [05:28<03:03,  2.03s/it]\n",
      "Loss: 0.01722:  64%|██████▍   | 160/250 [05:28<03:03,  2.03s/it]\n",
      "Loss: 0.01612:  64%|██████▍   | 160/250 [05:28<03:03,  2.03s/it]\n",
      "Loss: 0.01612:  64%|██████▍   | 161/250 [05:28<03:01,  2.04s/it]\n",
      "Loss: 0.01851:  64%|██████▍   | 161/250 [05:28<03:01,  2.04s/it]\n",
      "Loss: 0.01844:  64%|██████▍   | 161/250 [05:28<03:01,  2.04s/it]\n",
      "Loss: 0.02041:  64%|██████▍   | 161/250 [05:28<03:01,  2.04s/it]\n",
      "Loss: 0.01692:  64%|██████▍   | 161/250 [05:28<03:01,  2.04s/it]\n",
      "Loss: 0.01809:  64%|██████▍   | 161/250 [05:28<03:01,  2.04s/it]\n",
      "Loss: 0.01615:  64%|██████▍   | 161/250 [05:29<03:01,  2.04s/it]\n",
      "Loss: 0.01827:  64%|██████▍   | 161/250 [05:29<03:01,  2.04s/it]\n",
      "Loss: 0.02027:  64%|██████▍   | 161/250 [05:29<03:01,  2.04s/it]\n",
      "Loss: 0.01848:  64%|██████▍   | 161/250 [05:29<03:01,  2.04s/it]\n",
      "Loss: 0.02002:  64%|██████▍   | 161/250 [05:29<03:01,  2.04s/it]\n",
      "Loss: 0.01918:  64%|██████▍   | 161/250 [05:29<03:01,  2.04s/it]\n",
      "Loss: 0.02009:  64%|██████▍   | 161/250 [05:29<03:01,  2.04s/it]\n",
      "Loss: 0.01736:  64%|██████▍   | 161/250 [05:29<03:01,  2.04s/it]\n",
      "Loss: 0.01834:  64%|██████▍   | 161/250 [05:29<03:01,  2.04s/it]\n",
      "Loss: 0.01733:  64%|██████▍   | 161/250 [05:30<03:01,  2.04s/it]\n",
      "Loss: 0.01731:  64%|██████▍   | 161/250 [05:30<03:01,  2.04s/it]\n",
      "Loss: 0.01844:  64%|██████▍   | 161/250 [05:30<03:01,  2.04s/it]\n",
      "Loss: 0.01600:  64%|██████▍   | 161/250 [05:30<03:01,  2.04s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 60.21it/s]\u001b[A\n",
      "Loss: 0.01600:  65%|██████▍   | 162/250 [05:30<02:59,  2.03s/it]\n",
      "Loss: 0.01859:  65%|██████▍   | 162/250 [05:30<02:59,  2.03s/it]\n",
      "Loss: 0.01735:  65%|██████▍   | 162/250 [05:30<02:59,  2.03s/it]\n",
      "Loss: 0.02045:  65%|██████▍   | 162/250 [05:30<02:59,  2.03s/it]\n",
      "Loss: 0.01999:  65%|██████▍   | 162/250 [05:30<02:59,  2.03s/it]\n",
      "Loss: 0.01601:  65%|██████▍   | 162/250 [05:30<02:59,  2.03s/it]\n",
      "Loss: 0.01888:  65%|██████▍   | 162/250 [05:31<02:59,  2.03s/it]\n",
      "Loss: 0.01713:  65%|██████▍   | 162/250 [05:31<02:59,  2.03s/it]\n",
      "Loss: 0.01882:  65%|██████▍   | 162/250 [05:31<02:59,  2.03s/it]\n",
      "Loss: 0.01891:  65%|██████▍   | 162/250 [05:31<02:59,  2.03s/it]\n",
      "Loss: 0.01810:  65%|██████▍   | 162/250 [05:31<02:59,  2.03s/it]\n",
      "Loss: 0.02045:  65%|██████▍   | 162/250 [05:31<02:59,  2.03s/it]\n",
      "Loss: 0.01827:  65%|██████▍   | 162/250 [05:31<02:59,  2.03s/it]\n",
      "Loss: 0.01883:  65%|██████▍   | 162/250 [05:31<02:59,  2.03s/it]\n",
      "Loss: 0.02017:  65%|██████▍   | 162/250 [05:31<02:59,  2.03s/it]\n",
      "Loss: 0.01828:  65%|██████▍   | 162/250 [05:32<02:59,  2.03s/it]\n",
      "Loss: 0.01656:  65%|██████▍   | 162/250 [05:32<02:59,  2.03s/it]\n",
      "Loss: 0.01980:  65%|██████▍   | 162/250 [05:32<02:59,  2.03s/it]\n",
      "Loss: 0.01965:  65%|██████▍   | 162/250 [05:32<02:59,  2.03s/it]\n",
      "Loss: 0.01591:  65%|██████▍   | 162/250 [05:32<02:59,  2.03s/it]\n",
      "Loss: 0.01591:  65%|██████▌   | 163/250 [05:32<02:56,  2.03s/it]\n",
      "Loss: 0.01823:  65%|██████▌   | 163/250 [05:32<02:56,  2.03s/it]\n",
      "Loss: 0.01830:  65%|██████▌   | 163/250 [05:32<02:56,  2.03s/it]\n",
      "Loss: 0.01984:  65%|██████▌   | 163/250 [05:32<02:56,  2.03s/it]\n",
      "Loss: 0.01865:  65%|██████▌   | 163/250 [05:32<02:56,  2.03s/it]\n",
      "Loss: 0.01791:  65%|██████▌   | 163/250 [05:32<02:56,  2.03s/it]\n",
      "Loss: 0.01611:  65%|██████▌   | 163/250 [05:33<02:56,  2.03s/it]\n",
      "Loss: 0.01866:  65%|██████▌   | 163/250 [05:33<02:56,  2.03s/it]\n",
      "Loss: 0.02016:  65%|██████▌   | 163/250 [05:33<02:56,  2.03s/it]\n",
      "Loss: 0.01869:  65%|██████▌   | 163/250 [05:33<02:56,  2.03s/it]\n",
      "Loss: 0.01814:  65%|██████▌   | 163/250 [05:33<02:56,  2.03s/it]\n",
      "Loss: 0.02063:  65%|██████▌   | 163/250 [05:33<02:56,  2.03s/it]\n",
      "Loss: 0.01935:  65%|██████▌   | 163/250 [05:33<02:56,  2.03s/it]\n",
      "Loss: 0.01896:  65%|██████▌   | 163/250 [05:33<02:56,  2.03s/it]\n",
      "Loss: 0.01992:  65%|██████▌   | 163/250 [05:33<02:56,  2.03s/it]\n",
      "Loss: 0.01836:  65%|██████▌   | 163/250 [05:34<02:56,  2.03s/it]\n",
      "Loss: 0.01754:  65%|██████▌   | 163/250 [05:34<02:56,  2.03s/it]\n",
      "Loss: 0.01956:  65%|██████▌   | 163/250 [05:34<02:56,  2.03s/it]\n",
      "Loss: 0.01959:  65%|██████▌   | 163/250 [05:34<02:56,  2.03s/it]\n",
      "Loss: 0.01571:  65%|██████▌   | 163/250 [05:34<02:56,  2.03s/it]\n",
      "Loss: 0.01571:  66%|██████▌   | 164/250 [05:34<02:55,  2.04s/it]\n",
      "Loss: 0.01833:  66%|██████▌   | 164/250 [05:34<02:55,  2.04s/it]\n",
      "Loss: 0.01835:  66%|██████▌   | 164/250 [05:34<02:55,  2.04s/it]\n",
      "Loss: 0.01987:  66%|██████▌   | 164/250 [05:34<02:55,  2.04s/it]\n",
      "Loss: 0.01829:  66%|██████▌   | 164/250 [05:34<02:55,  2.04s/it]\n",
      "Loss: 0.01856:  66%|██████▌   | 164/250 [05:34<02:55,  2.04s/it]\n",
      "Loss: 0.01590:  66%|██████▌   | 164/250 [05:35<02:55,  2.04s/it]\n",
      "Loss: 0.01798:  66%|██████▌   | 164/250 [05:35<02:55,  2.04s/it]\n",
      "Loss: 0.01785:  66%|██████▌   | 164/250 [05:35<02:55,  2.04s/it]\n",
      "Loss: 0.01887:  66%|██████▌   | 164/250 [05:35<02:55,  2.04s/it]\n",
      "Loss: 0.01905:  66%|██████▌   | 164/250 [05:35<02:55,  2.04s/it]\n",
      "Loss: 0.01736:  66%|██████▌   | 164/250 [05:35<02:55,  2.04s/it]\n",
      "Loss: 0.01789:  66%|██████▌   | 164/250 [05:35<02:55,  2.04s/it]\n",
      "Loss: 0.01901:  66%|██████▌   | 164/250 [05:35<02:55,  2.04s/it]\n",
      "Loss: 0.01644:  66%|██████▌   | 164/250 [05:35<02:55,  2.04s/it]\n",
      "Loss: 0.01847:  66%|██████▌   | 164/250 [05:36<02:55,  2.04s/it]\n",
      "Loss: 0.01908:  66%|██████▌   | 164/250 [05:36<02:55,  2.04s/it]\n",
      "Loss: 0.01984:  66%|██████▌   | 164/250 [05:36<02:55,  2.04s/it]\n",
      "Loss: 0.01694:  66%|██████▌   | 164/250 [05:36<02:55,  2.04s/it]\n",
      "Loss: 0.01597:  66%|██████▌   | 164/250 [05:36<02:55,  2.04s/it]\n",
      "Loss: 0.01597:  66%|██████▌   | 165/250 [05:36<02:52,  2.03s/it]\n",
      "Loss: 0.01835:  66%|██████▌   | 165/250 [05:36<02:52,  2.03s/it]\n",
      "Loss: 0.01749:  66%|██████▌   | 165/250 [05:36<02:52,  2.03s/it]\n",
      "Loss: 0.02016:  66%|██████▌   | 165/250 [05:36<02:52,  2.03s/it]\n",
      "Loss: 0.01703:  66%|██████▌   | 165/250 [05:36<02:52,  2.03s/it]\n",
      "Loss: 0.01783:  66%|██████▌   | 165/250 [05:37<02:52,  2.03s/it]\n",
      "Loss: 0.01588:  66%|██████▌   | 165/250 [05:37<02:52,  2.03s/it]\n",
      "Loss: 0.01750:  66%|██████▌   | 165/250 [05:37<02:52,  2.03s/it]\n",
      "Loss: 0.01937:  66%|██████▌   | 165/250 [05:37<02:52,  2.03s/it]\n",
      "Loss: 0.01822:  66%|██████▌   | 165/250 [05:37<02:52,  2.03s/it]\n",
      "Loss: 0.01926:  66%|██████▌   | 165/250 [05:37<02:52,  2.03s/it]\n",
      "Loss: 0.01942:  66%|██████▌   | 165/250 [05:37<02:52,  2.03s/it]\n",
      "Loss: 0.01827:  66%|██████▌   | 165/250 [05:37<02:52,  2.03s/it]\n",
      "Loss: 0.01861:  66%|██████▌   | 165/250 [05:37<02:52,  2.03s/it]\n",
      "Loss: 0.01631:  66%|██████▌   | 165/250 [05:38<02:52,  2.03s/it]\n",
      "Loss: 0.01880:  66%|██████▌   | 165/250 [05:38<02:52,  2.03s/it]\n",
      "Loss: 0.01957:  66%|██████▌   | 165/250 [05:38<02:52,  2.03s/it]\n",
      "Loss: 0.02101:  66%|██████▌   | 165/250 [05:38<02:52,  2.03s/it]\n",
      "Loss: 0.01665:  66%|██████▌   | 165/250 [05:38<02:52,  2.03s/it]\n",
      "Loss: 0.01566:  66%|██████▌   | 165/250 [05:38<02:52,  2.03s/it]\n",
      "Loss: 0.01566:  66%|██████▋   | 166/250 [05:38<02:51,  2.04s/it]\n",
      "Loss: 0.01812:  66%|██████▋   | 166/250 [05:38<02:51,  2.04s/it]\n",
      "Loss: 0.01791:  66%|██████▋   | 166/250 [05:38<02:51,  2.04s/it]\n",
      "Loss: 0.01966:  66%|██████▋   | 166/250 [05:38<02:51,  2.04s/it]\n",
      "Loss: 0.01796:  66%|██████▋   | 166/250 [05:38<02:51,  2.04s/it]\n",
      "Loss: 0.01740:  66%|██████▋   | 166/250 [05:39<02:51,  2.04s/it]\n",
      "Loss: 0.01776:  66%|██████▋   | 166/250 [05:39<02:51,  2.04s/it]\n",
      "Loss: 0.01934:  66%|██████▋   | 166/250 [05:39<02:51,  2.04s/it]\n",
      "Loss: 0.01822:  66%|██████▋   | 166/250 [05:39<02:51,  2.04s/it]\n",
      "Loss: 0.01661:  66%|██████▋   | 166/250 [05:39<02:51,  2.04s/it]\n",
      "Loss: 0.01925:  66%|██████▋   | 166/250 [05:39<02:51,  2.04s/it]\n",
      "Loss: 0.02051:  66%|██████▋   | 166/250 [05:39<02:51,  2.04s/it]\n",
      "Loss: 0.01790:  66%|██████▋   | 166/250 [05:39<02:51,  2.04s/it]\n",
      "Loss: 0.01769:  66%|██████▋   | 166/250 [05:39<02:51,  2.04s/it]\n",
      "Loss: 0.01798:  66%|██████▋   | 166/250 [05:40<02:51,  2.04s/it]\n",
      "Loss: 0.01745:  66%|██████▋   | 166/250 [05:40<02:51,  2.04s/it]\n",
      "Loss: 0.01945:  66%|██████▋   | 166/250 [05:40<02:51,  2.04s/it]\n",
      "Loss: 0.01796:  66%|██████▋   | 166/250 [05:40<02:51,  2.04s/it]\n",
      "Loss: 0.01594:  66%|██████▋   | 166/250 [05:40<02:51,  2.04s/it]\n",
      "Loss: 0.01594:  67%|██████▋   | 167/250 [05:40<02:48,  2.03s/it]\n",
      "Loss: 0.01798:  67%|██████▋   | 167/250 [05:40<02:48,  2.03s/it]\n",
      "Loss: 0.01657:  67%|██████▋   | 167/250 [05:40<02:48,  2.03s/it]\n",
      "Loss: 0.02040:  67%|██████▋   | 167/250 [05:40<02:48,  2.03s/it]\n",
      "Loss: 0.01726:  67%|██████▋   | 167/250 [05:40<02:48,  2.03s/it]\n",
      "Loss: 0.01796:  67%|██████▋   | 167/250 [05:41<02:48,  2.03s/it]\n",
      "Loss: 0.01587:  67%|██████▋   | 167/250 [05:41<02:48,  2.03s/it]\n",
      "Loss: 0.01757:  67%|██████▋   | 167/250 [05:41<02:48,  2.03s/it]\n",
      "Loss: 0.01991:  67%|██████▋   | 167/250 [05:41<02:48,  2.03s/it]\n",
      "Loss: 0.01920:  67%|██████▋   | 167/250 [05:41<02:48,  2.03s/it]\n",
      "Loss: 0.01785:  67%|██████▋   | 167/250 [05:41<02:48,  2.03s/it]\n",
      "Loss: 0.02054:  67%|██████▋   | 167/250 [05:41<02:48,  2.03s/it]\n",
      "Loss: 0.01759:  67%|██████▋   | 167/250 [05:41<02:48,  2.03s/it]\n",
      "Loss: 0.01873:  67%|██████▋   | 167/250 [05:41<02:48,  2.03s/it]\n",
      "Loss: 0.01617:  67%|██████▋   | 167/250 [05:42<02:48,  2.03s/it]\n",
      "Loss: 0.01895:  67%|██████▋   | 167/250 [05:42<02:48,  2.03s/it]\n",
      "Loss: 0.01959:  67%|██████▋   | 167/250 [05:42<02:48,  2.03s/it]\n",
      "Loss: 0.02042:  67%|██████▋   | 167/250 [05:42<02:48,  2.03s/it]\n",
      "Loss: 0.01697:  67%|██████▋   | 167/250 [05:42<02:48,  2.03s/it]\n",
      "Loss: 0.01596:  67%|██████▋   | 167/250 [05:42<02:48,  2.03s/it]\n",
      "Loss: 0.01596:  67%|██████▋   | 168/250 [05:42<02:46,  2.03s/it]\n",
      "Loss: 0.01821:  67%|██████▋   | 168/250 [05:42<02:46,  2.03s/it]\n",
      "Loss: 0.01869:  67%|██████▋   | 168/250 [05:42<02:46,  2.03s/it]\n",
      "Loss: 0.02021:  67%|██████▋   | 168/250 [05:42<02:46,  2.03s/it]\n",
      "Loss: 0.01671:  67%|██████▋   | 168/250 [05:43<02:46,  2.03s/it]\n",
      "Loss: 0.01790:  67%|██████▋   | 168/250 [05:43<02:46,  2.03s/it]\n",
      "Loss: 0.01660:  67%|██████▋   | 168/250 [05:43<02:46,  2.03s/it]\n",
      "Loss: 0.01764:  67%|██████▋   | 168/250 [05:43<02:46,  2.03s/it]\n",
      "Loss: 0.01935:  67%|██████▋   | 168/250 [05:43<02:46,  2.03s/it]\n",
      "Loss: 0.01858:  67%|██████▋   | 168/250 [05:43<02:46,  2.03s/it]\n",
      "Loss: 0.01800:  67%|██████▋   | 168/250 [05:43<02:46,  2.03s/it]\n",
      "Loss: 0.02092:  67%|██████▋   | 168/250 [05:43<02:46,  2.03s/it]\n",
      "Loss: 0.01834:  67%|██████▋   | 168/250 [05:43<02:46,  2.03s/it]\n",
      "Loss: 0.01849:  67%|██████▋   | 168/250 [05:44<02:46,  2.03s/it]\n",
      "Loss: 0.01988:  67%|██████▋   | 168/250 [05:44<02:46,  2.03s/it]\n",
      "Loss: 0.01837:  67%|██████▋   | 168/250 [05:44<02:46,  2.03s/it]\n",
      "Loss: 0.01657:  67%|██████▋   | 168/250 [05:44<02:46,  2.03s/it]\n",
      "Loss: 0.01968:  67%|██████▋   | 168/250 [05:44<02:46,  2.03s/it]\n",
      "Loss: 0.01978:  67%|██████▋   | 168/250 [05:44<02:46,  2.03s/it]\n",
      "Loss: 0.01588:  67%|██████▋   | 168/250 [05:44<02:46,  2.03s/it]\n",
      "Loss: 0.01588:  68%|██████▊   | 169/250 [05:44<02:44,  2.04s/it]\n",
      "Loss: 0.01826:  68%|██████▊   | 169/250 [05:44<02:44,  2.04s/it]\n",
      "Loss: 0.01660:  68%|██████▊   | 169/250 [05:44<02:44,  2.04s/it]\n",
      "Loss: 0.02003:  68%|██████▊   | 169/250 [05:44<02:44,  2.04s/it]\n",
      "Loss: 0.01675:  68%|██████▊   | 169/250 [05:45<02:44,  2.04s/it]\n",
      "Loss: 0.01759:  68%|██████▊   | 169/250 [05:45<02:44,  2.04s/it]\n",
      "Loss: 0.01593:  68%|██████▊   | 169/250 [05:45<02:44,  2.04s/it]\n",
      "Loss: 0.01755:  68%|██████▊   | 169/250 [05:45<02:44,  2.04s/it]\n",
      "Loss: 0.01941:  68%|██████▊   | 169/250 [05:45<02:44,  2.04s/it]\n",
      "Loss: 0.01835:  68%|██████▊   | 169/250 [05:45<02:44,  2.04s/it]\n",
      "Loss: 0.01907:  68%|██████▊   | 169/250 [05:45<02:44,  2.04s/it]\n",
      "Loss: 0.02027:  68%|██████▊   | 169/250 [05:45<02:44,  2.04s/it]\n",
      "Loss: 0.01796:  68%|██████▊   | 169/250 [05:45<02:44,  2.04s/it]\n",
      "Loss: 0.01844:  68%|██████▊   | 169/250 [05:46<02:44,  2.04s/it]\n",
      "Loss: 0.01937:  68%|██████▊   | 169/250 [05:46<02:44,  2.04s/it]\n",
      "Loss: 0.01815:  68%|██████▊   | 169/250 [05:46<02:44,  2.04s/it]\n",
      "Loss: 0.01662:  68%|██████▊   | 169/250 [05:46<02:44,  2.04s/it]\n",
      "Loss: 0.01914:  68%|██████▊   | 169/250 [05:46<02:44,  2.04s/it]\n",
      "Loss: 0.01941:  68%|██████▊   | 169/250 [05:46<02:44,  2.04s/it]\n",
      "Loss: 0.01565:  68%|██████▊   | 169/250 [05:46<02:44,  2.04s/it]\n",
      "Loss: 0.01565:  68%|██████▊   | 170/250 [05:46<02:42,  2.03s/it]\n",
      "Loss: 0.01795:  68%|██████▊   | 170/250 [05:46<02:42,  2.03s/it]\n",
      "Loss: 0.01785:  68%|██████▊   | 170/250 [05:46<02:42,  2.03s/it]\n",
      "Loss: 0.02006:  68%|██████▊   | 170/250 [05:46<02:42,  2.03s/it]\n",
      "Loss: 0.01791:  68%|██████▊   | 170/250 [05:47<02:42,  2.03s/it]\n",
      "Loss: 0.01756:  68%|██████▊   | 170/250 [05:47<02:42,  2.03s/it]\n",
      "Loss: 0.01772:  68%|██████▊   | 170/250 [05:47<02:42,  2.03s/it]\n",
      "Loss: 0.01818:  68%|██████▊   | 170/250 [05:47<02:42,  2.03s/it]\n",
      "Loss: 0.01807:  68%|██████▊   | 170/250 [05:47<02:42,  2.03s/it]\n",
      "Loss: 0.01712:  68%|██████▊   | 170/250 [05:47<02:42,  2.03s/it]\n",
      "Loss: 0.01891:  68%|██████▊   | 170/250 [05:47<02:42,  2.03s/it]\n",
      "Loss: 0.02031:  68%|██████▊   | 170/250 [05:47<02:42,  2.03s/it]\n",
      "Loss: 0.01728:  68%|██████▊   | 170/250 [05:47<02:42,  2.03s/it]\n",
      "Loss: 0.01785:  68%|██████▊   | 170/250 [05:48<02:42,  2.03s/it]\n",
      "Loss: 0.01797:  68%|██████▊   | 170/250 [05:48<02:42,  2.03s/it]\n",
      "Loss: 0.01706:  68%|██████▊   | 170/250 [05:48<02:42,  2.03s/it]\n",
      "Loss: 0.01933:  68%|██████▊   | 170/250 [05:48<02:42,  2.03s/it]\n",
      "Loss: 0.01772:  68%|██████▊   | 170/250 [05:48<02:42,  2.03s/it]\n",
      "Loss: 0.01544:  68%|██████▊   | 170/250 [05:48<02:42,  2.03s/it]\n",
      "Loss: 0.01544:  68%|██████▊   | 171/250 [05:48<02:40,  2.03s/it]\n",
      "Loss: 0.01763:  68%|██████▊   | 171/250 [05:48<02:40,  2.03s/it]\n",
      "Loss: 0.01661:  68%|██████▊   | 171/250 [05:48<02:40,  2.03s/it]\n",
      "Loss: 0.02067:  68%|██████▊   | 171/250 [05:48<02:40,  2.03s/it]\n",
      "Loss: 0.01688:  68%|██████▊   | 171/250 [05:49<02:40,  2.03s/it]\n",
      "Loss: 0.01784:  68%|██████▊   | 171/250 [05:49<02:40,  2.03s/it]\n",
      "Loss: 0.01590:  68%|██████▊   | 171/250 [05:49<02:40,  2.03s/it]\n",
      "Loss: 0.01690:  68%|██████▊   | 171/250 [05:49<02:40,  2.03s/it]\n",
      "Loss: 0.01879:  68%|██████▊   | 171/250 [05:49<02:40,  2.03s/it]\n",
      "Loss: 0.01951:  68%|██████▊   | 171/250 [05:49<02:40,  2.03s/it]\n",
      "Loss: 0.01800:  68%|██████▊   | 171/250 [05:49<02:40,  2.03s/it]\n",
      "Loss: 0.01732:  68%|██████▊   | 171/250 [05:49<02:40,  2.03s/it]\n",
      "Loss: 0.02034:  68%|██████▊   | 171/250 [05:49<02:40,  2.03s/it]\n",
      "Loss: 0.01734:  68%|██████▊   | 171/250 [05:50<02:40,  2.03s/it]\n",
      "Loss: 0.01876:  68%|██████▊   | 171/250 [05:50<02:40,  2.03s/it]\n",
      "Loss: 0.01605:  68%|██████▊   | 171/250 [05:50<02:40,  2.03s/it]\n",
      "Loss: 0.01737:  68%|██████▊   | 171/250 [05:50<02:40,  2.03s/it]\n",
      "Loss: 0.01567:  68%|██████▊   | 171/250 [05:50<02:40,  2.03s/it]\n",
      "Loss: 0.01796:  68%|██████▊   | 171/250 [05:50<02:40,  2.03s/it]\n",
      "Loss: 0.01575:  68%|██████▊   | 171/250 [05:50<02:40,  2.03s/it]\n",
      "Loss: 0.01575:  69%|██████▉   | 172/250 [05:50<02:38,  2.04s/it]\n",
      "Loss: 0.01807:  69%|██████▉   | 172/250 [05:50<02:38,  2.04s/it]\n",
      "Loss: 0.01656:  69%|██████▉   | 172/250 [05:50<02:38,  2.04s/it]\n",
      "Loss: 0.01991:  69%|██████▉   | 172/250 [05:51<02:38,  2.04s/it]\n",
      "Loss: 0.01668:  69%|██████▉   | 172/250 [05:51<02:38,  2.04s/it]\n",
      "Loss: 0.01767:  69%|██████▉   | 172/250 [05:51<02:38,  2.04s/it]\n",
      "Loss: 0.01627:  69%|██████▉   | 172/250 [05:51<02:38,  2.04s/it]\n",
      "Loss: 0.01735:  69%|██████▉   | 172/250 [05:51<02:38,  2.04s/it]\n",
      "Loss: 0.01946:  69%|██████▉   | 172/250 [05:51<02:38,  2.04s/it]\n",
      "Loss: 0.01937:  69%|██████▉   | 172/250 [05:51<02:38,  2.04s/it]\n",
      "Loss: 0.01777:  69%|██████▉   | 172/250 [05:51<02:38,  2.04s/it]\n",
      "Loss: 0.02025:  69%|██████▉   | 172/250 [05:51<02:38,  2.04s/it]\n",
      "Loss: 0.01808:  69%|██████▉   | 172/250 [05:52<02:38,  2.04s/it]\n",
      "Loss: 0.01849:  69%|██████▉   | 172/250 [05:52<02:38,  2.04s/it]\n",
      "Loss: 0.01959:  69%|██████▉   | 172/250 [05:52<02:38,  2.04s/it]\n",
      "Loss: 0.01801:  69%|██████▉   | 172/250 [05:52<02:38,  2.04s/it]\n",
      "Loss: 0.01639:  69%|██████▉   | 172/250 [05:52<02:38,  2.04s/it]\n",
      "Loss: 0.01920:  69%|██████▉   | 172/250 [05:52<02:38,  2.04s/it]\n",
      "Loss: 0.01972:  69%|██████▉   | 172/250 [05:52<02:38,  2.04s/it]\n",
      "Loss: 0.01617:  69%|██████▉   | 172/250 [05:52<02:38,  2.04s/it]\n",
      "Loss: 0.01617:  69%|██████▉   | 173/250 [05:52<02:36,  2.04s/it]\n",
      "Loss: 0.01793:  69%|██████▉   | 173/250 [05:52<02:36,  2.04s/it]\n",
      "Loss: 0.01810:  69%|██████▉   | 173/250 [05:52<02:36,  2.04s/it]\n",
      "Loss: 0.01951:  69%|██████▉   | 173/250 [05:53<02:36,  2.04s/it]\n",
      "Loss: 0.01776:  69%|██████▉   | 173/250 [05:53<02:36,  2.04s/it]\n",
      "Loss: 0.01790:  69%|██████▉   | 173/250 [05:53<02:36,  2.04s/it]\n",
      "Loss: 0.01558:  69%|██████▉   | 173/250 [05:53<02:36,  2.04s/it]\n",
      "Loss: 0.01730:  69%|██████▉   | 173/250 [05:53<02:36,  2.04s/it]\n",
      "Loss: 0.01932:  69%|██████▉   | 173/250 [05:53<02:36,  2.04s/it]\n",
      "Loss: 0.01835:  69%|██████▉   | 173/250 [05:53<02:36,  2.04s/it]\n",
      "Loss: 0.01784:  69%|██████▉   | 173/250 [05:53<02:36,  2.04s/it]\n",
      "Loss: 0.02063:  69%|██████▉   | 173/250 [05:53<02:36,  2.04s/it]\n",
      "Loss: 0.01806:  69%|██████▉   | 173/250 [05:54<02:36,  2.04s/it]\n",
      "Loss: 0.01823:  69%|██████▉   | 173/250 [05:54<02:36,  2.04s/it]\n",
      "Loss: 0.01972:  69%|██████▉   | 173/250 [05:54<02:36,  2.04s/it]\n",
      "Loss: 0.01830:  69%|██████▉   | 173/250 [05:54<02:36,  2.04s/it]\n",
      "Loss: 0.01634:  69%|██████▉   | 173/250 [05:54<02:36,  2.04s/it]\n",
      "Loss: 0.01889:  69%|██████▉   | 173/250 [05:54<02:36,  2.04s/it]\n",
      "Loss: 0.01921:  69%|██████▉   | 173/250 [05:54<02:36,  2.04s/it]\n",
      "Loss: 0.01564:  69%|██████▉   | 173/250 [05:54<02:36,  2.04s/it]\n",
      "Loss: 0.01564:  70%|██████▉   | 174/250 [05:54<02:34,  2.04s/it]\n",
      "Loss: 0.01842:  70%|██████▉   | 174/250 [05:54<02:34,  2.04s/it]\n",
      "Loss: 0.01776:  70%|██████▉   | 174/250 [05:55<02:34,  2.04s/it]\n",
      "Loss: 0.01930:  70%|██████▉   | 174/250 [05:55<02:34,  2.04s/it]\n",
      "Loss: 0.01778:  70%|██████▉   | 174/250 [05:55<02:34,  2.04s/it]\n",
      "Loss: 0.01761:  70%|██████▉   | 174/250 [05:55<02:34,  2.04s/it]\n",
      "Loss: 0.01556:  70%|██████▉   | 174/250 [05:55<02:34,  2.04s/it]\n",
      "Loss: 0.01720:  70%|██████▉   | 174/250 [05:55<02:34,  2.04s/it]\n",
      "Loss: 0.01995:  70%|██████▉   | 174/250 [05:55<02:34,  2.04s/it]\n",
      "Loss: 0.01802:  70%|██████▉   | 174/250 [05:55<02:34,  2.04s/it]\n",
      "Loss: 0.01779:  70%|██████▉   | 174/250 [05:55<02:34,  2.04s/it]\n",
      "Loss: 0.02012:  70%|██████▉   | 174/250 [05:55<02:34,  2.04s/it]\n",
      "Loss: 0.01794:  70%|██████▉   | 174/250 [05:56<02:34,  2.04s/it]\n",
      "Loss: 0.01825:  70%|██████▉   | 174/250 [05:56<02:34,  2.04s/it]\n",
      "Loss: 0.01949:  70%|██████▉   | 174/250 [05:56<02:34,  2.04s/it]\n",
      "Loss: 0.01841:  70%|██████▉   | 174/250 [05:56<02:34,  2.04s/it]\n",
      "Loss: 0.01617:  70%|██████▉   | 174/250 [05:56<02:34,  2.04s/it]\n",
      "Loss: 0.01941:  70%|██████▉   | 174/250 [05:56<02:34,  2.04s/it]\n",
      "Loss: 0.01927:  70%|██████▉   | 174/250 [05:56<02:34,  2.04s/it]\n",
      "Loss: 0.01568:  70%|██████▉   | 174/250 [05:56<02:34,  2.04s/it]\n",
      "Loss: 0.01568:  70%|███████   | 175/250 [05:56<02:32,  2.04s/it]\n",
      "Loss: 0.01819:  70%|███████   | 175/250 [05:56<02:32,  2.04s/it]\n",
      "Loss: 0.01637:  70%|███████   | 175/250 [05:57<02:32,  2.04s/it]\n",
      "Loss: 0.01625:  70%|███████   | 175/250 [05:57<02:32,  2.04s/it]\n",
      "Loss: 0.01922:  70%|███████   | 175/250 [05:57<02:32,  2.04s/it]\n",
      "Loss: 0.01577:  70%|███████   | 175/250 [05:57<02:32,  2.04s/it]\n",
      "Loss: 0.01847:  70%|███████   | 175/250 [05:57<02:32,  2.04s/it]\n",
      "Loss: 0.01632:  70%|███████   | 175/250 [05:57<02:32,  2.04s/it]\n",
      "Loss: 0.01753:  70%|███████   | 175/250 [05:57<02:32,  2.04s/it]\n",
      "Loss: 0.01935:  70%|███████   | 175/250 [05:57<02:32,  2.04s/it]\n",
      "Loss: 0.01801:  70%|███████   | 175/250 [05:57<02:32,  2.04s/it]\n",
      "Loss: 0.02032:  70%|███████   | 175/250 [05:58<02:32,  2.04s/it]\n",
      "Loss: 0.01769:  70%|███████   | 175/250 [05:58<02:32,  2.04s/it]\n",
      "Loss: 0.01857:  70%|███████   | 175/250 [05:58<02:32,  2.04s/it]\n",
      "Loss: 0.01950:  70%|███████   | 175/250 [05:58<02:32,  2.04s/it]\n",
      "Loss: 0.01894:  70%|███████   | 175/250 [05:58<02:32,  2.04s/it]\n",
      "Loss: 0.01650:  70%|███████   | 175/250 [05:58<02:32,  2.04s/it]\n",
      "Loss: 0.01878:  70%|███████   | 175/250 [05:58<02:32,  2.04s/it]\n",
      "Loss: 0.01946:  70%|███████   | 175/250 [05:58<02:32,  2.04s/it]\n",
      "Loss: 0.01594:  70%|███████   | 175/250 [05:58<02:32,  2.04s/it]\n",
      "Loss: 0.01594:  70%|███████   | 176/250 [05:58<02:30,  2.03s/it]\n",
      "Loss: 0.01760:  70%|███████   | 176/250 [05:58<02:30,  2.03s/it]\n",
      "Loss: 0.01641:  70%|███████   | 176/250 [05:59<02:30,  2.03s/it]\n",
      "Loss: 0.02035:  70%|███████   | 176/250 [05:59<02:30,  2.03s/it]\n",
      "Loss: 0.01750:  70%|███████   | 176/250 [05:59<02:30,  2.03s/it]\n",
      "Loss: 0.01776:  70%|███████   | 176/250 [05:59<02:30,  2.03s/it]\n",
      "Loss: 0.01881:  70%|███████   | 176/250 [05:59<02:30,  2.03s/it]\n",
      "Loss: 0.01685:  70%|███████   | 176/250 [05:59<02:30,  2.03s/it]\n",
      "Loss: 0.01770:  70%|███████   | 176/250 [05:59<02:30,  2.03s/it]\n",
      "Loss: 0.01853:  70%|███████   | 176/250 [05:59<02:30,  2.03s/it]\n",
      "Loss: 0.01803:  70%|███████   | 176/250 [05:59<02:30,  2.03s/it]\n",
      "Loss: 0.02025:  70%|███████   | 176/250 [06:00<02:30,  2.03s/it]\n",
      "Loss: 0.01775:  70%|███████   | 176/250 [06:00<02:30,  2.03s/it]\n",
      "Loss: 0.01865:  70%|███████   | 176/250 [06:00<02:30,  2.03s/it]\n",
      "Loss: 0.01922:  70%|███████   | 176/250 [06:00<02:30,  2.03s/it]\n",
      "Loss: 0.01781:  70%|███████   | 176/250 [06:00<02:30,  2.03s/it]\n",
      "Loss: 0.01619:  70%|███████   | 176/250 [06:00<02:30,  2.03s/it]\n",
      "Loss: 0.01873:  70%|███████   | 176/250 [06:00<02:30,  2.03s/it]\n",
      "Loss: 0.01903:  70%|███████   | 176/250 [06:00<02:30,  2.03s/it]\n",
      "Loss: 0.01533:  70%|███████   | 176/250 [06:00<02:30,  2.03s/it]\n",
      "Loss: 0.01533:  71%|███████   | 177/250 [06:00<02:27,  2.02s/it]\n",
      "Loss: 0.01834:  71%|███████   | 177/250 [06:00<02:27,  2.02s/it]\n",
      "Loss: 0.01762:  71%|███████   | 177/250 [06:01<02:27,  2.02s/it]\n",
      "Loss: 0.01943:  71%|███████   | 177/250 [06:01<02:27,  2.02s/it]\n",
      "Loss: 0.01777:  71%|███████   | 177/250 [06:01<02:27,  2.02s/it]\n",
      "Loss: 0.01774:  71%|███████   | 177/250 [06:01<02:27,  2.02s/it]\n",
      "Loss: 0.01785:  71%|███████   | 177/250 [06:01<02:27,  2.02s/it]\n",
      "Loss: 0.01812:  71%|███████   | 177/250 [06:01<02:27,  2.02s/it]\n",
      "Loss: 0.01770:  71%|███████   | 177/250 [06:01<02:27,  2.02s/it]\n",
      "Loss: 0.01610:  71%|███████   | 177/250 [06:01<02:27,  2.02s/it]\n",
      "Loss: 0.01881:  71%|███████   | 177/250 [06:01<02:27,  2.02s/it]\n",
      "Loss: 0.01953:  71%|███████   | 177/250 [06:02<02:27,  2.02s/it]\n",
      "Loss: 0.01717:  71%|███████   | 177/250 [06:02<02:27,  2.02s/it]\n",
      "Loss: 0.01768:  71%|███████   | 177/250 [06:02<02:27,  2.02s/it]\n",
      "Loss: 0.01774:  71%|███████   | 177/250 [06:02<02:27,  2.02s/it]\n",
      "Loss: 0.01681:  71%|███████   | 177/250 [06:02<02:27,  2.02s/it]\n",
      "Loss: 0.01899:  71%|███████   | 177/250 [06:02<02:27,  2.02s/it]\n",
      "Loss: 0.01806:  71%|███████   | 177/250 [06:02<02:27,  2.02s/it]\n",
      "Loss: 0.01548:  71%|███████   | 177/250 [06:02<02:27,  2.02s/it]\n",
      "Loss: 0.01548:  71%|███████   | 178/250 [06:02<02:25,  2.02s/it]\n",
      "Loss: 0.01781:  71%|███████   | 178/250 [06:02<02:25,  2.02s/it]\n",
      "Loss: 0.01759:  71%|███████   | 178/250 [06:03<02:25,  2.02s/it]\n",
      "Loss: 0.01980:  71%|███████   | 178/250 [06:03<02:25,  2.02s/it]\n",
      "Loss: 0.01666:  71%|███████   | 178/250 [06:03<02:25,  2.02s/it]\n",
      "Loss: 0.01758:  71%|███████   | 178/250 [06:03<02:25,  2.02s/it]\n",
      "Loss: 0.01578:  71%|███████   | 178/250 [06:03<02:25,  2.02s/it]\n",
      "Loss: 0.01799:  71%|███████   | 178/250 [06:03<02:25,  2.02s/it]\n",
      "Loss: 0.01903:  71%|███████   | 178/250 [06:03<02:25,  2.02s/it]\n",
      "Loss: 0.01816:  71%|███████   | 178/250 [06:03<02:25,  2.02s/it]\n",
      "Loss: 0.01888:  71%|███████   | 178/250 [06:03<02:25,  2.02s/it]\n",
      "Loss: 0.01943:  71%|███████   | 178/250 [06:04<02:25,  2.02s/it]\n",
      "Loss: 0.01869:  71%|███████   | 178/250 [06:04<02:25,  2.02s/it]\n",
      "Loss: 0.01678:  71%|███████   | 178/250 [06:04<02:25,  2.02s/it]\n",
      "Loss: 0.01745:  71%|███████   | 178/250 [06:04<02:25,  2.02s/it]\n",
      "Loss: 0.01647:  71%|███████   | 178/250 [06:04<02:25,  2.02s/it]\n",
      "Loss: 0.01636:  71%|███████   | 178/250 [06:04<02:25,  2.02s/it]\n",
      "Loss: 0.01801:  71%|███████   | 178/250 [06:04<02:25,  2.02s/it]\n",
      "Loss: 0.01655:  71%|███████   | 178/250 [06:04<02:25,  2.02s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 63.46it/s]\u001b[A\n",
      "Loss: 0.01655:  72%|███████▏  | 179/250 [06:04<02:23,  2.02s/it]\n",
      "Loss: 0.01794:  72%|███████▏  | 179/250 [06:04<02:23,  2.02s/it]\n",
      "Loss: 0.01747:  72%|███████▏  | 179/250 [06:05<02:23,  2.02s/it]\n",
      "Loss: 0.01918:  72%|███████▏  | 179/250 [06:05<02:23,  2.02s/it]\n",
      "Loss: 0.01771:  72%|███████▏  | 179/250 [06:05<02:23,  2.02s/it]\n",
      "Loss: 0.01700:  72%|███████▏  | 179/250 [06:05<02:23,  2.02s/it]\n",
      "Loss: 0.01742:  72%|███████▏  | 179/250 [06:05<02:23,  2.02s/it]\n",
      "Loss: 0.01780:  72%|███████▏  | 179/250 [06:05<02:23,  2.02s/it]\n",
      "Loss: 0.01813:  72%|███████▏  | 179/250 [06:05<02:23,  2.02s/it]\n",
      "Loss: 0.01775:  72%|███████▏  | 179/250 [06:05<02:23,  2.02s/it]\n",
      "Loss: 0.01762:  72%|███████▏  | 179/250 [06:05<02:23,  2.02s/it]\n",
      "Loss: 0.02011:  72%|███████▏  | 179/250 [06:06<02:23,  2.02s/it]\n",
      "Loss: 0.01759:  72%|███████▏  | 179/250 [06:06<02:23,  2.02s/it]\n",
      "Loss: 0.01829:  72%|███████▏  | 179/250 [06:06<02:23,  2.02s/it]\n",
      "Loss: 0.01931:  72%|███████▏  | 179/250 [06:06<02:23,  2.02s/it]\n",
      "Loss: 0.01753:  72%|███████▏  | 179/250 [06:06<02:23,  2.02s/it]\n",
      "Loss: 0.01626:  72%|███████▏  | 179/250 [06:06<02:23,  2.02s/it]\n",
      "Loss: 0.01865:  72%|███████▏  | 179/250 [06:06<02:23,  2.02s/it]\n",
      "Loss: 0.01915:  72%|███████▏  | 179/250 [06:06<02:23,  2.02s/it]\n",
      "Loss: 0.01565:  72%|███████▏  | 179/250 [06:06<02:23,  2.02s/it]\n",
      "Loss: 0.01565:  72%|███████▏  | 180/250 [06:06<02:21,  2.02s/it]\n",
      "Loss: 0.01783:  72%|███████▏  | 180/250 [06:06<02:21,  2.02s/it]\n",
      "Loss: 0.01801:  72%|███████▏  | 180/250 [06:07<02:21,  2.02s/it]\n",
      "Loss: 0.01918:  72%|███████▏  | 180/250 [06:07<02:21,  2.02s/it]\n",
      "Loss: 0.01784:  72%|███████▏  | 180/250 [06:07<02:21,  2.02s/it]\n",
      "Loss: 0.01729:  72%|███████▏  | 180/250 [06:07<02:21,  2.02s/it]\n",
      "Loss: 0.01790:  72%|███████▏  | 180/250 [06:07<02:21,  2.02s/it]\n",
      "Loss: 0.01809:  72%|███████▏  | 180/250 [06:07<02:21,  2.02s/it]\n",
      "Loss: 0.01916:  72%|███████▏  | 180/250 [06:07<02:21,  2.02s/it]\n",
      "Loss: 0.01802:  72%|███████▏  | 180/250 [06:07<02:21,  2.02s/it]\n",
      "Loss: 0.01912:  72%|███████▏  | 180/250 [06:07<02:21,  2.02s/it]\n",
      "Loss: 0.01871:  72%|███████▏  | 180/250 [06:08<02:21,  2.02s/it]\n",
      "Loss: 0.01893:  72%|███████▏  | 180/250 [06:08<02:21,  2.02s/it]\n",
      "Loss: 0.01705:  72%|███████▏  | 180/250 [06:08<02:21,  2.02s/it]\n",
      "Loss: 0.01748:  72%|███████▏  | 180/250 [06:08<02:21,  2.02s/it]\n",
      "Loss: 0.01611:  72%|███████▏  | 180/250 [06:08<02:21,  2.02s/it]\n",
      "Loss: 0.01679:  72%|███████▏  | 180/250 [06:08<02:21,  2.02s/it]\n",
      "Loss: 0.01795:  72%|███████▏  | 180/250 [06:08<02:21,  2.02s/it]\n",
      "Loss: 0.01569:  72%|███████▏  | 180/250 [06:08<02:21,  2.02s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.31it/s]\u001b[A\n",
      "Loss: 0.01569:  72%|███████▏  | 181/250 [06:08<02:19,  2.02s/it]\n",
      "Loss: 0.01834:  72%|███████▏  | 181/250 [06:09<02:19,  2.02s/it]\n",
      "Loss: 0.01743:  72%|███████▏  | 181/250 [06:09<02:19,  2.02s/it]\n",
      "Loss: 0.01925:  72%|███████▏  | 181/250 [06:09<02:19,  2.02s/it]\n",
      "Loss: 0.01746:  72%|███████▏  | 181/250 [06:09<02:19,  2.02s/it]\n",
      "Loss: 0.01742:  72%|███████▏  | 181/250 [06:09<02:19,  2.02s/it]\n",
      "Loss: 0.01753:  72%|███████▏  | 181/250 [06:09<02:19,  2.02s/it]\n",
      "Loss: 0.01803:  72%|███████▏  | 181/250 [06:09<02:19,  2.02s/it]\n",
      "Loss: 0.01794:  72%|███████▏  | 181/250 [06:09<02:19,  2.02s/it]\n",
      "Loss: 0.01647:  72%|███████▏  | 181/250 [06:09<02:19,  2.02s/it]\n",
      "Loss: 0.01911:  72%|███████▏  | 181/250 [06:10<02:19,  2.02s/it]\n",
      "Loss: 0.02013:  72%|███████▏  | 181/250 [06:10<02:19,  2.02s/it]\n",
      "Loss: 0.01686:  72%|███████▏  | 181/250 [06:10<02:19,  2.02s/it]\n",
      "Loss: 0.01748:  72%|███████▏  | 181/250 [06:10<02:19,  2.02s/it]\n",
      "Loss: 0.01769:  72%|███████▏  | 181/250 [06:10<02:19,  2.02s/it]\n",
      "Loss: 0.01731:  72%|███████▏  | 181/250 [06:10<02:19,  2.02s/it]\n",
      "Loss: 0.01854:  72%|███████▏  | 181/250 [06:10<02:19,  2.02s/it]\n",
      "Loss: 0.01794:  72%|███████▏  | 181/250 [06:10<02:19,  2.02s/it]\n",
      "Loss: 0.01563:  72%|███████▏  | 181/250 [06:10<02:19,  2.02s/it]\n",
      "Loss: 0.01563:  73%|███████▎  | 182/250 [06:10<02:17,  2.02s/it]\n",
      "Loss: 0.01777:  73%|███████▎  | 182/250 [06:11<02:17,  2.02s/it]\n",
      "Loss: 0.01776:  73%|███████▎  | 182/250 [06:11<02:17,  2.02s/it]\n",
      "Loss: 0.01936:  73%|███████▎  | 182/250 [06:11<02:17,  2.02s/it]\n",
      "Loss: 0.01768:  73%|███████▎  | 182/250 [06:11<02:17,  2.02s/it]\n",
      "Loss: 0.01723:  73%|███████▎  | 182/250 [06:11<02:17,  2.02s/it]\n",
      "Loss: 0.01598:  73%|███████▎  | 182/250 [06:11<02:17,  2.02s/it]\n",
      "Loss: 0.01692:  73%|███████▎  | 182/250 [06:11<02:17,  2.02s/it]\n",
      "Loss: 0.01757:  73%|███████▎  | 182/250 [06:11<02:17,  2.02s/it]\n",
      "Loss: 0.01914:  73%|███████▎  | 182/250 [06:11<02:17,  2.02s/it]\n",
      "Loss: 0.01793:  73%|███████▎  | 182/250 [06:12<02:17,  2.02s/it]\n",
      "Loss: 0.01999:  73%|███████▎  | 182/250 [06:12<02:17,  2.02s/it]\n",
      "Loss: 0.01833:  73%|███████▎  | 182/250 [06:12<02:17,  2.02s/it]\n",
      "Loss: 0.01849:  73%|███████▎  | 182/250 [06:12<02:17,  2.02s/it]\n",
      "Loss: 0.01996:  73%|███████▎  | 182/250 [06:12<02:17,  2.02s/it]\n",
      "Loss: 0.01770:  73%|███████▎  | 182/250 [06:12<02:17,  2.02s/it]\n",
      "Loss: 0.01651:  73%|███████▎  | 182/250 [06:12<02:17,  2.02s/it]\n",
      "Loss: 0.01908:  73%|███████▎  | 182/250 [06:12<02:17,  2.02s/it]\n",
      "Loss: 0.01907:  73%|███████▎  | 182/250 [06:12<02:17,  2.02s/it]\n",
      "Loss: 0.01557:  73%|███████▎  | 182/250 [06:12<02:17,  2.02s/it]\n",
      "Loss: 0.01557:  73%|███████▎  | 183/250 [06:12<02:15,  2.02s/it]\n",
      "Loss: 0.01755:  73%|███████▎  | 183/250 [06:13<02:15,  2.02s/it]\n",
      "Loss: 0.01768:  73%|███████▎  | 183/250 [06:13<02:15,  2.02s/it]\n",
      "Loss: 0.01995:  73%|███████▎  | 183/250 [06:13<02:15,  2.02s/it]\n",
      "Loss: 0.01661:  73%|███████▎  | 183/250 [06:13<02:15,  2.02s/it]\n",
      "Loss: 0.01752:  73%|███████▎  | 183/250 [06:13<02:15,  2.02s/it]\n",
      "Loss: 0.01830:  73%|███████▎  | 183/250 [06:13<02:15,  2.02s/it]\n",
      "Loss: 0.01626:  73%|███████▎  | 183/250 [06:13<02:15,  2.02s/it]\n",
      "Loss: 0.01776:  73%|███████▎  | 183/250 [06:13<02:15,  2.02s/it]\n",
      "Loss: 0.01840:  73%|███████▎  | 183/250 [06:13<02:15,  2.02s/it]\n",
      "Loss: 0.01801:  73%|███████▎  | 183/250 [06:14<02:15,  2.02s/it]\n",
      "Loss: 0.01720:  73%|███████▎  | 183/250 [06:14<02:15,  2.02s/it]\n",
      "Loss: 0.01793:  73%|███████▎  | 183/250 [06:14<02:15,  2.02s/it]\n",
      "Loss: 0.01815:  73%|███████▎  | 183/250 [06:14<02:15,  2.02s/it]\n",
      "Loss: 0.01543:  73%|███████▎  | 183/250 [06:14<02:15,  2.02s/it]\n",
      "Loss: 0.01798:  73%|███████▎  | 183/250 [06:14<02:15,  2.02s/it]\n",
      "Loss: 0.01878:  73%|███████▎  | 183/250 [06:14<02:15,  2.02s/it]\n",
      "Loss: 0.02038:  73%|███████▎  | 183/250 [06:14<02:15,  2.02s/it]\n",
      "Loss: 0.01635:  73%|███████▎  | 183/250 [06:14<02:15,  2.02s/it]\n",
      "Loss: 0.01544:  73%|███████▎  | 183/250 [06:14<02:15,  2.02s/it]\n",
      "Loss: 0.01544:  74%|███████▎  | 184/250 [06:14<02:13,  2.02s/it]\n",
      "Loss: 0.01783:  74%|███████▎  | 184/250 [06:15<02:13,  2.02s/it]\n",
      "Loss: 0.01752:  74%|███████▎  | 184/250 [06:15<02:13,  2.02s/it]\n",
      "Loss: 0.01951:  74%|███████▎  | 184/250 [06:15<02:13,  2.02s/it]\n",
      "Loss: 0.01662:  74%|███████▎  | 184/250 [06:15<02:13,  2.02s/it]\n",
      "Loss: 0.01729:  74%|███████▎  | 184/250 [06:15<02:13,  2.02s/it]\n",
      "Loss: 0.01556:  74%|███████▎  | 184/250 [06:15<02:13,  2.02s/it]\n",
      "Loss: 0.01758:  74%|███████▎  | 184/250 [06:15<02:13,  2.02s/it]\n",
      "Loss: 0.01933:  74%|███████▎  | 184/250 [06:15<02:13,  2.02s/it]\n",
      "Loss: 0.01757:  74%|███████▎  | 184/250 [06:15<02:13,  2.02s/it]\n",
      "Loss: 0.01883:  74%|███████▎  | 184/250 [06:16<02:13,  2.02s/it]\n",
      "Loss: 0.01871:  74%|███████▎  | 184/250 [06:16<02:13,  2.02s/it]\n",
      "Loss: 0.01959:  74%|███████▎  | 184/250 [06:16<02:13,  2.02s/it]\n",
      "Loss: 0.01699:  74%|███████▎  | 184/250 [06:16<02:13,  2.02s/it]\n",
      "Loss: 0.01737:  74%|███████▎  | 184/250 [06:16<02:13,  2.02s/it]\n",
      "Loss: 0.01671:  74%|███████▎  | 184/250 [06:16<02:13,  2.02s/it]\n",
      "Loss: 0.01641:  74%|███████▎  | 184/250 [06:16<02:13,  2.02s/it]\n",
      "Loss: 0.01725:  74%|███████▎  | 184/250 [06:16<02:13,  2.02s/it]\n",
      "Loss: 0.01547:  74%|███████▎  | 184/250 [06:16<02:13,  2.02s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 61.25it/s]\u001b[A\n",
      "Loss: 0.01547:  74%|███████▍  | 185/250 [06:16<02:11,  2.02s/it]\n",
      "Loss: 0.01751:  74%|███████▍  | 185/250 [06:17<02:11,  2.02s/it]\n",
      "Loss: 0.01775:  74%|███████▍  | 185/250 [06:17<02:11,  2.02s/it]\n",
      "Loss: 0.02029:  74%|███████▍  | 185/250 [06:17<02:11,  2.02s/it]\n",
      "Loss: 0.01647:  74%|███████▍  | 185/250 [06:17<02:11,  2.02s/it]\n",
      "Loss: 0.01735:  74%|███████▍  | 185/250 [06:17<02:11,  2.02s/it]\n",
      "Loss: 0.01578:  74%|███████▍  | 185/250 [06:17<02:11,  2.02s/it]\n",
      "Loss: 0.01723:  74%|███████▍  | 185/250 [06:17<02:11,  2.02s/it]\n",
      "Loss: 0.01748:  74%|███████▍  | 185/250 [06:17<02:11,  2.02s/it]\n",
      "Loss: 0.01836:  74%|███████▍  | 185/250 [06:17<02:11,  2.02s/it]\n",
      "Loss: 0.01817:  74%|███████▍  | 185/250 [06:18<02:11,  2.02s/it]\n",
      "Loss: 0.01723:  74%|███████▍  | 185/250 [06:18<02:11,  2.02s/it]\n",
      "Loss: 0.01747:  74%|███████▍  | 185/250 [06:18<02:11,  2.02s/it]\n",
      "Loss: 0.01842:  74%|███████▍  | 185/250 [06:18<02:11,  2.02s/it]\n",
      "Loss: 0.01590:  74%|███████▍  | 185/250 [06:18<02:11,  2.02s/it]\n",
      "Loss: 0.01818:  74%|███████▍  | 185/250 [06:18<02:11,  2.02s/it]\n",
      "Loss: 0.01953:  74%|███████▍  | 185/250 [06:18<02:11,  2.02s/it]\n",
      "Loss: 0.01984:  74%|███████▍  | 185/250 [06:18<02:11,  2.02s/it]\n",
      "Loss: 0.01733:  74%|███████▍  | 185/250 [06:18<02:11,  2.02s/it]\n",
      "Loss: 0.01584:  74%|███████▍  | 185/250 [06:18<02:11,  2.02s/it]\n",
      "Loss: 0.01584:  74%|███████▍  | 186/250 [06:19<02:09,  2.02s/it]\n",
      "Loss: 0.01744:  74%|███████▍  | 186/250 [06:19<02:09,  2.02s/it]\n",
      "Loss: 0.01870:  74%|███████▍  | 186/250 [06:19<02:09,  2.02s/it]\n",
      "Loss: 0.01948:  74%|███████▍  | 186/250 [06:19<02:09,  2.02s/it]\n",
      "Loss: 0.01744:  74%|███████▍  | 186/250 [06:19<02:09,  2.02s/it]\n",
      "Loss: 0.01729:  74%|███████▍  | 186/250 [06:19<02:09,  2.02s/it]\n",
      "Loss: 0.01749:  74%|███████▍  | 186/250 [06:19<02:09,  2.02s/it]\n",
      "Loss: 0.01798:  74%|███████▍  | 186/250 [06:19<02:09,  2.02s/it]\n",
      "Loss: 0.01897:  74%|███████▍  | 186/250 [06:19<02:09,  2.02s/it]\n",
      "Loss: 0.01845:  74%|███████▍  | 186/250 [06:19<02:09,  2.02s/it]\n",
      "Loss: 0.01773:  74%|███████▍  | 186/250 [06:20<02:09,  2.02s/it]\n",
      "Loss: 0.01973:  74%|███████▍  | 186/250 [06:20<02:09,  2.02s/it]\n",
      "Loss: 0.01773:  74%|███████▍  | 186/250 [06:20<02:09,  2.02s/it]\n",
      "Loss: 0.01807:  74%|███████▍  | 186/250 [06:20<02:09,  2.02s/it]\n",
      "Loss: 0.01918:  74%|███████▍  | 186/250 [06:20<02:09,  2.02s/it]\n",
      "Loss: 0.01792:  74%|███████▍  | 186/250 [06:20<02:09,  2.02s/it]\n",
      "Loss: 0.01622:  74%|███████▍  | 186/250 [06:20<02:09,  2.02s/it]\n",
      "Loss: 0.01899:  74%|███████▍  | 186/250 [06:20<02:09,  2.02s/it]\n",
      "Loss: 0.01927:  74%|███████▍  | 186/250 [06:20<02:09,  2.02s/it]\n",
      "Loss: 0.01557:  74%|███████▍  | 186/250 [06:21<02:09,  2.02s/it]\n",
      "Loss: 0.01557:  75%|███████▍  | 187/250 [06:21<02:07,  2.02s/it]\n",
      "Loss: 0.01816:  75%|███████▍  | 187/250 [06:21<02:07,  2.02s/it]\n",
      "Loss: 0.01736:  75%|███████▍  | 187/250 [06:21<02:07,  2.02s/it]\n",
      "Loss: 0.01968:  75%|███████▍  | 187/250 [06:21<02:07,  2.02s/it]\n",
      "Loss: 0.01764:  75%|███████▍  | 187/250 [06:21<02:07,  2.02s/it]\n",
      "Loss: 0.01706:  75%|███████▍  | 187/250 [06:21<02:07,  2.02s/it]\n",
      "Loss: 0.01766:  75%|███████▍  | 187/250 [06:21<02:07,  2.02s/it]\n",
      "Loss: 0.01711:  75%|███████▍  | 187/250 [06:21<02:07,  2.02s/it]\n",
      "Loss: 0.01743:  75%|███████▍  | 187/250 [06:21<02:07,  2.02s/it]\n",
      "Loss: 0.01832:  75%|███████▍  | 187/250 [06:21<02:07,  2.02s/it]\n",
      "Loss: 0.01771:  75%|███████▍  | 187/250 [06:22<02:07,  2.02s/it]\n",
      "Loss: 0.02037:  75%|███████▍  | 187/250 [06:22<02:07,  2.02s/it]\n",
      "Loss: 0.01757:  75%|███████▍  | 187/250 [06:22<02:07,  2.02s/it]\n",
      "Loss: 0.01882:  75%|███████▍  | 187/250 [06:22<02:07,  2.02s/it]\n",
      "Loss: 0.01922:  75%|███████▍  | 187/250 [06:22<02:07,  2.02s/it]\n",
      "Loss: 0.01799:  75%|███████▍  | 187/250 [06:22<02:07,  2.02s/it]\n",
      "Loss: 0.01852:  75%|███████▍  | 187/250 [06:22<02:07,  2.02s/it]\n",
      "Loss: 0.01507:  75%|███████▍  | 187/250 [06:22<02:07,  2.02s/it]\n",
      "Loss: 0.01788:  75%|███████▍  | 187/250 [06:23<02:07,  2.02s/it]\n",
      "Loss: 0.01552:  75%|███████▍  | 187/250 [06:23<02:07,  2.02s/it]\n",
      "Loss: 0.01552:  75%|███████▌  | 188/250 [06:23<02:05,  2.03s/it]\n",
      "Loss: 0.01765:  75%|███████▌  | 188/250 [06:23<02:05,  2.03s/it]\n",
      "Loss: 0.01752:  75%|███████▌  | 188/250 [06:23<02:05,  2.03s/it]\n",
      "Loss: 0.01950:  75%|███████▌  | 188/250 [06:23<02:05,  2.03s/it]\n",
      "Loss: 0.01787:  75%|███████▌  | 188/250 [06:23<02:05,  2.03s/it]\n",
      "Loss: 0.01734:  75%|███████▌  | 188/250 [06:23<02:05,  2.03s/it]\n",
      "Loss: 0.01745:  75%|███████▌  | 188/250 [06:23<02:05,  2.03s/it]\n",
      "Loss: 0.01673:  75%|███████▌  | 188/250 [06:23<02:05,  2.03s/it]\n",
      "Loss: 0.01901:  75%|███████▌  | 188/250 [06:23<02:05,  2.03s/it]\n",
      "Loss: 0.01798:  75%|███████▌  | 188/250 [06:24<02:05,  2.03s/it]\n",
      "Loss: 0.01892:  75%|███████▌  | 188/250 [06:24<02:05,  2.03s/it]\n",
      "Loss: 0.01860:  75%|███████▌  | 188/250 [06:24<02:05,  2.03s/it]\n",
      "Loss: 0.01904:  75%|███████▌  | 188/250 [06:24<02:05,  2.03s/it]\n",
      "Loss: 0.01679:  75%|███████▌  | 188/250 [06:24<02:05,  2.03s/it]\n",
      "Loss: 0.01726:  75%|███████▌  | 188/250 [06:24<02:05,  2.03s/it]\n",
      "Loss: 0.01635:  75%|███████▌  | 188/250 [06:24<02:05,  2.03s/it]\n",
      "Loss: 0.01661:  75%|███████▌  | 188/250 [06:24<02:05,  2.03s/it]\n",
      "Loss: 0.01767:  75%|███████▌  | 188/250 [06:24<02:05,  2.03s/it]\n",
      "Loss: 0.01584:  75%|███████▌  | 188/250 [06:25<02:05,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 62.24it/s]\u001b[A\n",
      "Loss: 0.01584:  76%|███████▌  | 189/250 [06:25<02:03,  2.02s/it]\n",
      "Loss: 0.01778:  76%|███████▌  | 189/250 [06:25<02:03,  2.02s/it]\n",
      "Loss: 0.01752:  76%|███████▌  | 189/250 [06:25<02:03,  2.02s/it]\n",
      "Loss: 0.01919:  76%|███████▌  | 189/250 [06:25<02:03,  2.02s/it]\n",
      "Loss: 0.01773:  76%|███████▌  | 189/250 [06:25<02:03,  2.02s/it]\n",
      "Loss: 0.01738:  76%|███████▌  | 189/250 [06:25<02:03,  2.02s/it]\n",
      "Loss: 0.01562:  76%|███████▌  | 189/250 [06:25<02:03,  2.02s/it]\n",
      "Loss: 0.01704:  76%|███████▌  | 189/250 [06:25<02:03,  2.02s/it]\n",
      "Loss: 0.01749:  76%|███████▌  | 189/250 [06:25<02:03,  2.02s/it]\n",
      "Loss: 0.01837:  76%|███████▌  | 189/250 [06:26<02:03,  2.02s/it]\n",
      "Loss: 0.01746:  76%|███████▌  | 189/250 [06:26<02:03,  2.02s/it]\n",
      "Loss: 0.02035:  76%|███████▌  | 189/250 [06:26<02:03,  2.02s/it]\n",
      "Loss: 0.01739:  76%|███████▌  | 189/250 [06:26<02:03,  2.02s/it]\n",
      "Loss: 0.01840:  76%|███████▌  | 189/250 [06:26<02:03,  2.02s/it]\n",
      "Loss: 0.01970:  76%|███████▌  | 189/250 [06:26<02:03,  2.02s/it]\n",
      "Loss: 0.01738:  76%|███████▌  | 189/250 [06:26<02:03,  2.02s/it]\n",
      "Loss: 0.01624:  76%|███████▌  | 189/250 [06:26<02:03,  2.02s/it]\n",
      "Loss: 0.01957:  76%|███████▌  | 189/250 [06:26<02:03,  2.02s/it]\n",
      "Loss: 0.01879:  76%|███████▌  | 189/250 [06:27<02:03,  2.02s/it]\n",
      "Loss: 0.01526:  76%|███████▌  | 189/250 [06:27<02:03,  2.02s/it]\n",
      "Loss: 0.01526:  76%|███████▌  | 190/250 [06:27<02:01,  2.03s/it]\n",
      "Loss: 0.01773:  76%|███████▌  | 190/250 [06:27<02:01,  2.03s/it]\n",
      "Loss: 0.01731:  76%|███████▌  | 190/250 [06:27<02:01,  2.03s/it]\n",
      "Loss: 0.01969:  76%|███████▌  | 190/250 [06:27<02:01,  2.03s/it]\n",
      "Loss: 0.01745:  76%|███████▌  | 190/250 [06:27<02:01,  2.03s/it]\n",
      "Loss: 0.01747:  76%|███████▌  | 190/250 [06:27<02:01,  2.03s/it]\n",
      "Loss: 0.01537:  76%|███████▌  | 190/250 [06:27<02:01,  2.03s/it]\n",
      "Loss: 0.01712:  76%|███████▌  | 190/250 [06:27<02:01,  2.03s/it]\n",
      "Loss: 0.01873:  76%|███████▌  | 190/250 [06:27<02:01,  2.03s/it]\n",
      "Loss: 0.01777:  76%|███████▌  | 190/250 [06:28<02:01,  2.03s/it]\n",
      "Loss: 0.01907:  76%|███████▌  | 190/250 [06:28<02:01,  2.03s/it]\n",
      "Loss: 0.01898:  76%|███████▌  | 190/250 [06:28<02:01,  2.03s/it]\n",
      "Loss: 0.01871:  76%|███████▌  | 190/250 [06:28<02:01,  2.03s/it]\n",
      "Loss: 0.01696:  76%|███████▌  | 190/250 [06:28<02:01,  2.03s/it]\n",
      "Loss: 0.01763:  76%|███████▌  | 190/250 [06:28<02:01,  2.03s/it]\n",
      "Loss: 0.01651:  76%|███████▌  | 190/250 [06:28<02:01,  2.03s/it]\n",
      "Loss: 0.01726:  76%|███████▌  | 190/250 [06:28<02:01,  2.03s/it]\n",
      "Loss: 0.01748:  76%|███████▌  | 190/250 [06:29<02:01,  2.03s/it]\n",
      "Loss: 0.01581:  76%|███████▌  | 190/250 [06:29<02:01,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 62.72it/s]\u001b[A\n",
      "Loss: 0.01581:  76%|███████▋  | 191/250 [06:29<01:59,  2.03s/it]\n",
      "Loss: 0.01735:  76%|███████▋  | 191/250 [06:29<01:59,  2.03s/it]\n",
      "Loss: 0.01624:  76%|███████▋  | 191/250 [06:29<01:59,  2.03s/it]\n",
      "Loss: 0.01970:  76%|███████▋  | 191/250 [06:29<01:59,  2.03s/it]\n",
      "Loss: 0.01900:  76%|███████▋  | 191/250 [06:29<01:59,  2.03s/it]\n",
      "Loss: 0.01806:  76%|███████▋  | 191/250 [06:29<01:59,  2.03s/it]\n",
      "Loss: 0.01825:  76%|███████▋  | 191/250 [06:29<01:59,  2.03s/it]\n",
      "Loss: 0.01982:  76%|███████▋  | 191/250 [06:29<01:59,  2.03s/it]\n",
      "Loss: 0.01769:  76%|███████▋  | 191/250 [06:30<01:59,  2.03s/it]\n",
      "Loss: 0.01836:  76%|███████▋  | 191/250 [06:30<01:59,  2.03s/it]\n",
      "Loss: 0.01803:  76%|███████▋  | 191/250 [06:30<01:59,  2.03s/it]\n",
      "Loss: 0.01687:  76%|███████▋  | 191/250 [06:30<01:59,  2.03s/it]\n",
      "Loss: 0.01837:  76%|███████▋  | 191/250 [06:30<01:59,  2.03s/it]\n",
      "Loss: 0.01839:  76%|███████▋  | 191/250 [06:30<01:59,  2.03s/it]\n",
      "Loss: 0.01543:  76%|███████▋  | 191/250 [06:30<01:59,  2.03s/it]\n",
      "Loss: 0.01800:  76%|███████▋  | 191/250 [06:30<01:59,  2.03s/it]\n",
      "Loss: 0.01852:  76%|███████▋  | 191/250 [06:30<01:59,  2.03s/it]\n",
      "Loss: 0.01963:  76%|███████▋  | 191/250 [06:31<01:59,  2.03s/it]\n",
      "Loss: 0.01625:  76%|███████▋  | 191/250 [06:31<01:59,  2.03s/it]\n",
      "Loss: 0.01551:  76%|███████▋  | 191/250 [06:31<01:59,  2.03s/it]\n",
      "Loss: 0.01551:  77%|███████▋  | 192/250 [06:31<01:57,  2.03s/it]\n",
      "Loss: 0.01766:  77%|███████▋  | 192/250 [06:31<01:57,  2.03s/it]\n",
      "Loss: 0.01748:  77%|███████▋  | 192/250 [06:31<01:57,  2.03s/it]\n",
      "Loss: 0.01921:  77%|███████▋  | 192/250 [06:31<01:57,  2.03s/it]\n",
      "Loss: 0.01755:  77%|███████▋  | 192/250 [06:31<01:57,  2.03s/it]\n",
      "Loss: 0.01752:  77%|███████▋  | 192/250 [06:31<01:57,  2.03s/it]\n",
      "Loss: 0.01595:  77%|███████▋  | 192/250 [06:31<01:57,  2.03s/it]\n",
      "Loss: 0.01599:  77%|███████▋  | 192/250 [06:31<01:57,  2.03s/it]\n",
      "Loss: 0.01789:  77%|███████▋  | 192/250 [06:32<01:57,  2.03s/it]\n",
      "Loss: 0.01841:  77%|███████▋  | 192/250 [06:32<01:57,  2.03s/it]\n",
      "Loss: 0.01765:  77%|███████▋  | 192/250 [06:32<01:57,  2.03s/it]\n",
      "Loss: 0.01649:  77%|███████▋  | 192/250 [06:32<01:57,  2.03s/it]\n",
      "Loss: 0.01945:  77%|███████▋  | 192/250 [06:32<01:57,  2.03s/it]\n",
      "Loss: 0.01683:  77%|███████▋  | 192/250 [06:32<01:57,  2.03s/it]\n",
      "Loss: 0.01922:  77%|███████▋  | 192/250 [06:32<01:57,  2.03s/it]\n",
      "Loss: 0.01621:  77%|███████▋  | 192/250 [06:32<01:57,  2.03s/it]\n",
      "Loss: 0.01700:  77%|███████▋  | 192/250 [06:32<01:57,  2.03s/it]\n",
      "Loss: 0.01586:  77%|███████▋  | 192/250 [06:33<01:57,  2.03s/it]\n",
      "Loss: 0.01809:  77%|███████▋  | 192/250 [06:33<01:57,  2.03s/it]\n",
      "Loss: 0.01570:  77%|███████▋  | 192/250 [06:33<01:57,  2.03s/it]\n",
      "Loss: 0.01570:  77%|███████▋  | 193/250 [06:33<01:55,  2.03s/it]\n",
      "Loss: 0.01743:  77%|███████▋  | 193/250 [06:33<01:55,  2.03s/it]\n",
      "Loss: 0.01674:  77%|███████▋  | 193/250 [06:33<01:55,  2.03s/it]\n",
      "Loss: 0.01599:  77%|███████▋  | 193/250 [06:33<01:55,  2.03s/it]\n",
      "Loss: 0.01902:  77%|███████▋  | 193/250 [06:33<01:55,  2.03s/it]\n",
      "Loss: 0.01562:  77%|███████▋  | 193/250 [06:33<01:55,  2.03s/it]\n",
      "Loss: 0.01868:  77%|███████▋  | 193/250 [06:33<01:55,  2.03s/it]\n",
      "Loss: 0.01949:  77%|███████▋  | 193/250 [06:33<01:55,  2.03s/it]\n",
      "Loss: 0.01918:  77%|███████▋  | 193/250 [06:34<01:55,  2.03s/it]\n",
      "Loss: 0.01730:  77%|███████▋  | 193/250 [06:34<01:55,  2.03s/it]\n",
      "Loss: 0.01922:  77%|███████▋  | 193/250 [06:34<01:55,  2.03s/it]\n",
      "Loss: 0.01733:  77%|███████▋  | 193/250 [06:34<01:55,  2.03s/it]\n",
      "Loss: 0.01583:  77%|███████▋  | 193/250 [06:34<01:55,  2.03s/it]\n",
      "Loss: 0.02079:  77%|███████▋  | 193/250 [06:34<01:55,  2.03s/it]\n",
      "Loss: 0.01816:  77%|███████▋  | 193/250 [06:34<01:55,  2.03s/it]\n",
      "Loss: 0.01722:  77%|███████▋  | 193/250 [06:34<01:55,  2.03s/it]\n",
      "Loss: 0.02064:  77%|███████▋  | 193/250 [06:34<01:55,  2.03s/it]\n",
      "Loss: 0.01568:  77%|███████▋  | 193/250 [06:35<01:55,  2.03s/it]\n",
      "Loss: 0.01738:  77%|███████▋  | 193/250 [06:35<01:55,  2.03s/it]\n",
      "Loss: 0.01547:  77%|███████▋  | 193/250 [06:35<01:55,  2.03s/it]\n",
      "Loss: 0.01547:  78%|███████▊  | 194/250 [06:35<01:53,  2.03s/it]\n",
      "Loss: 0.01788:  78%|███████▊  | 194/250 [06:35<01:53,  2.03s/it]\n",
      "Loss: 0.01732:  78%|███████▊  | 194/250 [06:35<01:53,  2.03s/it]\n",
      "Loss: 0.01954:  78%|███████▊  | 194/250 [06:35<01:53,  2.03s/it]\n",
      "Loss: 0.01895:  78%|███████▊  | 194/250 [06:35<01:53,  2.03s/it]\n",
      "Loss: 0.01820:  78%|███████▊  | 194/250 [06:35<01:53,  2.03s/it]\n",
      "Loss: 0.01843:  78%|███████▊  | 194/250 [06:35<01:53,  2.03s/it]\n",
      "Loss: 0.01979:  78%|███████▊  | 194/250 [06:36<01:53,  2.03s/it]\n",
      "Loss: 0.01799:  78%|███████▊  | 194/250 [06:36<01:53,  2.03s/it]\n",
      "Loss: 0.01903:  78%|███████▊  | 194/250 [06:36<01:53,  2.03s/it]\n",
      "Loss: 0.01765:  78%|███████▊  | 194/250 [06:36<01:53,  2.03s/it]\n",
      "Loss: 0.01658:  78%|███████▊  | 194/250 [06:36<01:53,  2.03s/it]\n",
      "Loss: 0.01966:  78%|███████▊  | 194/250 [06:36<01:53,  2.03s/it]\n",
      "Loss: 0.01673:  78%|███████▊  | 194/250 [06:36<01:53,  2.03s/it]\n",
      "Loss: 0.01884:  78%|███████▊  | 194/250 [06:36<01:53,  2.03s/it]\n",
      "Loss: 0.01568:  78%|███████▊  | 194/250 [06:36<01:53,  2.03s/it]\n",
      "Loss: 0.01711:  78%|███████▊  | 194/250 [06:37<01:53,  2.03s/it]\n",
      "Loss: 0.01545:  78%|███████▊  | 194/250 [06:37<01:53,  2.03s/it]\n",
      "Loss: 0.01791:  78%|███████▊  | 194/250 [06:37<01:53,  2.03s/it]\n",
      "Loss: 0.01543:  78%|███████▊  | 194/250 [06:37<01:53,  2.03s/it]\n",
      "Loss: 0.01543:  78%|███████▊  | 195/250 [06:37<01:52,  2.04s/it]\n",
      "Loss: 0.01731:  78%|███████▊  | 195/250 [06:37<01:52,  2.04s/it]\n",
      "Loss: 0.01660:  78%|███████▊  | 195/250 [06:37<01:52,  2.04s/it]\n",
      "Loss: 0.02062:  78%|███████▊  | 195/250 [06:37<01:52,  2.04s/it]\n",
      "Loss: 0.01674:  78%|███████▊  | 195/250 [06:37<01:52,  2.04s/it]\n",
      "Loss: 0.01542:  78%|███████▊  | 195/250 [06:37<01:52,  2.04s/it]\n",
      "Loss: 0.01816:  78%|███████▊  | 195/250 [06:37<01:52,  2.04s/it]\n",
      "Loss: 0.01649:  78%|███████▊  | 195/250 [06:38<01:52,  2.04s/it]\n",
      "Loss: 0.01764:  78%|███████▊  | 195/250 [06:38<01:52,  2.04s/it]\n",
      "Loss: 0.01841:  78%|███████▊  | 195/250 [06:38<01:52,  2.04s/it]\n",
      "Loss: 0.01776:  78%|███████▊  | 195/250 [06:38<01:52,  2.04s/it]\n",
      "Loss: 0.01986:  78%|███████▊  | 195/250 [06:38<01:52,  2.04s/it]\n",
      "Loss: 0.01752:  78%|███████▊  | 195/250 [06:38<01:52,  2.04s/it]\n",
      "Loss: 0.01841:  78%|███████▊  | 195/250 [06:38<01:52,  2.04s/it]\n",
      "Loss: 0.01991:  78%|███████▊  | 195/250 [06:38<01:52,  2.04s/it]\n",
      "Loss: 0.01746:  78%|███████▊  | 195/250 [06:38<01:52,  2.04s/it]\n",
      "Loss: 0.01633:  78%|███████▊  | 195/250 [06:39<01:52,  2.04s/it]\n",
      "Loss: 0.01873:  78%|███████▊  | 195/250 [06:39<01:52,  2.04s/it]\n",
      "Loss: 0.01936:  78%|███████▊  | 195/250 [06:39<01:52,  2.04s/it]\n",
      "Loss: 0.01564:  78%|███████▊  | 195/250 [06:39<01:52,  2.04s/it]\n",
      "Loss: 0.01564:  78%|███████▊  | 196/250 [06:39<01:50,  2.04s/it]\n",
      "Loss: 0.01770:  78%|███████▊  | 196/250 [06:39<01:50,  2.04s/it]\n",
      "Loss: 0.01750:  78%|███████▊  | 196/250 [06:39<01:50,  2.04s/it]\n",
      "Loss: 0.01930:  78%|███████▊  | 196/250 [06:39<01:50,  2.04s/it]\n",
      "Loss: 0.01696:  78%|███████▊  | 196/250 [06:39<01:50,  2.04s/it]\n",
      "Loss: 0.01732:  78%|███████▊  | 196/250 [06:39<01:50,  2.04s/it]\n",
      "Loss: 0.01536:  78%|███████▊  | 196/250 [06:40<01:50,  2.04s/it]\n",
      "Loss: 0.01712:  78%|███████▊  | 196/250 [06:40<01:50,  2.04s/it]\n",
      "Loss: 0.01874:  78%|███████▊  | 196/250 [06:40<01:50,  2.04s/it]\n",
      "Loss: 0.01784:  78%|███████▊  | 196/250 [06:40<01:50,  2.04s/it]\n",
      "Loss: 0.01727:  78%|███████▊  | 196/250 [06:40<01:50,  2.04s/it]\n",
      "Loss: 0.01997:  78%|███████▊  | 196/250 [06:40<01:50,  2.04s/it]\n",
      "Loss: 0.01712:  78%|███████▊  | 196/250 [06:40<01:50,  2.04s/it]\n",
      "Loss: 0.01819:  78%|███████▊  | 196/250 [06:40<01:50,  2.04s/it]\n",
      "Loss: 0.01610:  78%|███████▊  | 196/250 [06:40<01:50,  2.04s/it]\n",
      "Loss: 0.01776:  78%|███████▊  | 196/250 [06:41<01:50,  2.04s/it]\n",
      "Loss: 0.01738:  78%|███████▊  | 196/250 [06:41<01:50,  2.04s/it]\n",
      "Loss: 0.01503:  78%|███████▊  | 196/250 [06:41<01:50,  2.04s/it]\n",
      "Loss: 0.01782:  78%|███████▊  | 196/250 [06:41<01:50,  2.04s/it]\n",
      "Loss: 0.01543:  78%|███████▊  | 196/250 [06:41<01:50,  2.04s/it]\n",
      "Loss: 0.01543:  79%|███████▉  | 197/250 [06:41<01:48,  2.05s/it]\n",
      "Loss: 0.01734:  79%|███████▉  | 197/250 [06:41<01:48,  2.05s/it]\n",
      "Loss: 0.01640:  79%|███████▉  | 197/250 [06:41<01:48,  2.05s/it]\n",
      "Loss: 0.01606:  79%|███████▉  | 197/250 [06:41<01:48,  2.05s/it]\n",
      "Loss: 0.01909:  79%|███████▉  | 197/250 [06:41<01:48,  2.05s/it]\n",
      "Loss: 0.01797:  79%|███████▉  | 197/250 [06:41<01:48,  2.05s/it]\n",
      "Loss: 0.01727:  79%|███████▉  | 197/250 [06:42<01:48,  2.05s/it]\n",
      "Loss: 0.02114:  79%|███████▉  | 197/250 [06:42<01:48,  2.05s/it]\n",
      "Loss: 0.01998:  79%|███████▉  | 197/250 [06:42<01:48,  2.05s/it]\n",
      "Loss: 0.01768:  79%|███████▉  | 197/250 [06:42<01:48,  2.05s/it]\n",
      "Loss: 0.01888:  79%|███████▉  | 197/250 [06:42<01:48,  2.05s/it]\n",
      "Loss: 0.01859:  79%|███████▉  | 197/250 [06:42<01:48,  2.05s/it]\n",
      "Loss: 0.01926:  79%|███████▉  | 197/250 [06:42<01:48,  2.05s/it]\n",
      "Loss: 0.01796:  79%|███████▉  | 197/250 [06:42<01:48,  2.05s/it]\n",
      "Loss: 0.01619:  79%|███████▉  | 197/250 [06:42<01:48,  2.05s/it]\n",
      "Loss: 0.01690:  79%|███████▉  | 197/250 [06:43<01:48,  2.05s/it]\n",
      "Loss: 0.01751:  79%|███████▉  | 197/250 [06:43<01:48,  2.05s/it]\n",
      "Loss: 0.01902:  79%|███████▉  | 197/250 [06:43<01:48,  2.05s/it]\n",
      "Loss: 0.01641:  79%|███████▉  | 197/250 [06:43<01:48,  2.05s/it]\n",
      "Loss: 0.01532:  79%|███████▉  | 197/250 [06:43<01:48,  2.05s/it]\n",
      "Loss: 0.01532:  79%|███████▉  | 198/250 [06:43<01:46,  2.05s/it]\n",
      "Loss: 0.01761:  79%|███████▉  | 198/250 [06:43<01:46,  2.05s/it]\n",
      "Loss: 0.01744:  79%|███████▉  | 198/250 [06:43<01:46,  2.05s/it]\n",
      "Loss: 0.01879:  79%|███████▉  | 198/250 [06:43<01:46,  2.05s/it]\n",
      "Loss: 0.01791:  79%|███████▉  | 198/250 [06:43<01:46,  2.05s/it]\n",
      "Loss: 0.01703:  79%|███████▉  | 198/250 [06:44<01:46,  2.05s/it]\n",
      "Loss: 0.01792:  79%|███████▉  | 198/250 [06:44<01:46,  2.05s/it]\n",
      "Loss: 0.01686:  79%|███████▉  | 198/250 [06:44<01:46,  2.05s/it]\n",
      "Loss: 0.01864:  79%|███████▉  | 198/250 [06:44<01:46,  2.05s/it]\n",
      "Loss: 0.01779:  79%|███████▉  | 198/250 [06:44<01:46,  2.05s/it]\n",
      "Loss: 0.01758:  79%|███████▉  | 198/250 [06:44<01:46,  2.05s/it]\n",
      "Loss: 0.01991:  79%|███████▉  | 198/250 [06:44<01:46,  2.05s/it]\n",
      "Loss: 0.01751:  79%|███████▉  | 198/250 [06:44<01:46,  2.05s/it]\n",
      "Loss: 0.01785:  79%|███████▉  | 198/250 [06:44<01:46,  2.05s/it]\n",
      "Loss: 0.01867:  79%|███████▉  | 198/250 [06:45<01:46,  2.05s/it]\n",
      "Loss: 0.01725:  79%|███████▉  | 198/250 [06:45<01:46,  2.05s/it]\n",
      "Loss: 0.01603:  79%|███████▉  | 198/250 [06:45<01:46,  2.05s/it]\n",
      "Loss: 0.01884:  79%|███████▉  | 198/250 [06:45<01:46,  2.05s/it]\n",
      "Loss: 0.01883:  79%|███████▉  | 198/250 [06:45<01:46,  2.05s/it]\n",
      "Loss: 0.01537:  79%|███████▉  | 198/250 [06:45<01:46,  2.05s/it]\n",
      "Loss: 0.01537:  80%|███████▉  | 199/250 [06:45<01:44,  2.04s/it]\n",
      "Loss: 0.01767:  80%|███████▉  | 199/250 [06:45<01:44,  2.04s/it]\n",
      "Loss: 0.01757:  80%|███████▉  | 199/250 [06:45<01:44,  2.04s/it]\n",
      "Loss: 0.02006:  80%|███████▉  | 199/250 [06:45<01:44,  2.04s/it]\n",
      "Loss: 0.01744:  80%|███████▉  | 199/250 [06:45<01:44,  2.04s/it]\n",
      "Loss: 0.01704:  80%|███████▉  | 199/250 [06:46<01:44,  2.04s/it]\n",
      "Loss: 0.01745:  80%|███████▉  | 199/250 [06:46<01:44,  2.04s/it]\n",
      "Loss: 0.01806:  80%|███████▉  | 199/250 [06:46<01:44,  2.04s/it]\n",
      "Loss: 0.01860:  80%|███████▉  | 199/250 [06:46<01:44,  2.04s/it]\n",
      "Loss: 0.01737:  80%|███████▉  | 199/250 [06:46<01:44,  2.04s/it]\n",
      "Loss: 0.01762:  80%|███████▉  | 199/250 [06:46<01:44,  2.04s/it]\n",
      "Loss: 0.01704:  80%|███████▉  | 199/250 [06:46<01:44,  2.04s/it]\n",
      "Loss: 0.01732:  80%|███████▉  | 199/250 [06:46<01:44,  2.04s/it]\n",
      "Loss: 0.01850:  80%|███████▉  | 199/250 [06:46<01:44,  2.04s/it]\n",
      "Loss: 0.01579:  80%|███████▉  | 199/250 [06:47<01:44,  2.04s/it]\n",
      "Loss: 0.01784:  80%|███████▉  | 199/250 [06:47<01:44,  2.04s/it]\n",
      "Loss: 0.01871:  80%|███████▉  | 199/250 [06:47<01:44,  2.04s/it]\n",
      "Loss: 0.01965:  80%|███████▉  | 199/250 [06:47<01:44,  2.04s/it]\n",
      "Loss: 0.01621:  80%|███████▉  | 199/250 [06:47<01:44,  2.04s/it]\n",
      "Loss: 0.01545:  80%|███████▉  | 199/250 [06:47<01:44,  2.04s/it]\n",
      "Loss: 0.01545:  80%|████████  | 200/250 [06:47<01:42,  2.04s/it]\n",
      "Loss: 0.01743:  80%|████████  | 200/250 [06:47<01:42,  2.04s/it]\n",
      "Loss: 0.01729:  80%|████████  | 200/250 [06:47<01:42,  2.04s/it]\n",
      "Loss: 0.01900:  80%|████████  | 200/250 [06:47<01:42,  2.04s/it]\n",
      "Loss: 0.01766:  80%|████████  | 200/250 [06:47<01:42,  2.04s/it]\n",
      "Loss: 0.01719:  80%|████████  | 200/250 [06:48<01:42,  2.04s/it]\n",
      "Loss: 0.01739:  80%|████████  | 200/250 [06:48<01:42,  2.04s/it]\n",
      "Loss: 0.01795:  80%|████████  | 200/250 [06:48<01:42,  2.04s/it]\n",
      "Loss: 0.01785:  80%|████████  | 200/250 [06:48<01:42,  2.04s/it]\n",
      "Loss: 0.01598:  80%|████████  | 200/250 [06:48<01:42,  2.04s/it]\n",
      "Loss: 0.01821:  80%|████████  | 200/250 [06:48<01:42,  2.04s/it]\n",
      "Loss: 0.01981:  80%|████████  | 200/250 [06:48<01:42,  2.04s/it]\n",
      "Loss: 0.01692:  80%|████████  | 200/250 [06:48<01:42,  2.04s/it]\n",
      "Loss: 0.01682:  80%|████████  | 200/250 [06:48<01:42,  2.04s/it]\n",
      "Loss: 0.01682:  80%|████████  | 200/250 [06:49<01:42,  2.04s/it]\n",
      "Loss: 0.01591:  80%|████████  | 200/250 [06:49<01:42,  2.04s/it]\n",
      "Loss: 0.01600:  80%|████████  | 200/250 [06:49<01:42,  2.04s/it]\n",
      "Loss: 0.01794:  80%|████████  | 200/250 [06:49<01:42,  2.04s/it]\n",
      "Loss: 0.01539:  80%|████████  | 200/250 [06:49<01:42,  2.04s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.51it/s]\u001b[A\n",
      "Loss: 0.01539:  80%|████████  | 201/250 [06:49<01:40,  2.04s/it]\n",
      "Loss: 0.01874:  80%|████████  | 201/250 [06:49<01:40,  2.04s/it]\n",
      "Loss: 0.01748:  80%|████████  | 201/250 [06:49<01:40,  2.04s/it]\n",
      "Loss: 0.01905:  80%|████████  | 201/250 [06:49<01:40,  2.04s/it]\n",
      "Loss: 0.01739:  80%|████████  | 201/250 [06:50<01:40,  2.04s/it]\n",
      "Loss: 0.01708:  80%|████████  | 201/250 [06:50<01:40,  2.04s/it]\n",
      "Loss: 0.01727:  80%|████████  | 201/250 [06:50<01:40,  2.04s/it]\n",
      "Loss: 0.01792:  80%|████████  | 201/250 [06:50<01:40,  2.04s/it]\n",
      "Loss: 0.01788:  80%|████████  | 201/250 [06:50<01:40,  2.04s/it]\n",
      "Loss: 0.01615:  80%|████████  | 201/250 [06:50<01:40,  2.04s/it]\n",
      "Loss: 0.01894:  80%|████████  | 201/250 [06:50<01:40,  2.04s/it]\n",
      "Loss: 0.01956:  80%|████████  | 201/250 [06:50<01:40,  2.04s/it]\n",
      "Loss: 0.01665:  80%|████████  | 201/250 [06:50<01:40,  2.04s/it]\n",
      "Loss: 0.01731:  80%|████████  | 201/250 [06:51<01:40,  2.04s/it]\n",
      "Loss: 0.01789:  80%|████████  | 201/250 [06:51<01:40,  2.04s/it]\n",
      "Loss: 0.01754:  80%|████████  | 201/250 [06:51<01:40,  2.04s/it]\n",
      "Loss: 0.01852:  80%|████████  | 201/250 [06:51<01:40,  2.04s/it]\n",
      "Loss: 0.01751:  80%|████████  | 201/250 [06:51<01:40,  2.04s/it]\n",
      "Loss: 0.01538:  80%|████████  | 201/250 [06:51<01:40,  2.04s/it]\n",
      "Loss: 0.01538:  81%|████████  | 202/250 [06:51<01:37,  2.04s/it]\n",
      "Loss: 0.01729:  81%|████████  | 202/250 [06:51<01:37,  2.04s/it]\n",
      "Loss: 0.01603:  81%|████████  | 202/250 [06:51<01:37,  2.04s/it]\n",
      "Loss: 0.01623:  81%|████████  | 202/250 [06:51<01:37,  2.04s/it]\n",
      "Loss: 0.01703:  81%|████████  | 202/250 [06:52<01:37,  2.04s/it]\n",
      "Loss: 0.01787:  81%|████████  | 202/250 [06:52<01:37,  2.04s/it]\n",
      "Loss: 0.01814:  81%|████████  | 202/250 [06:52<01:37,  2.04s/it]\n",
      "Loss: 0.01928:  81%|████████  | 202/250 [06:52<01:37,  2.04s/it]\n",
      "Loss: 0.01763:  81%|████████  | 202/250 [06:52<01:37,  2.04s/it]\n",
      "Loss: 0.01911:  81%|████████  | 202/250 [06:52<01:37,  2.04s/it]\n",
      "Loss: 0.01890:  81%|████████  | 202/250 [06:52<01:37,  2.04s/it]\n",
      "Loss: 0.01690:  81%|████████  | 202/250 [06:52<01:37,  2.04s/it]\n",
      "Loss: 0.01575:  81%|████████  | 202/250 [06:52<01:37,  2.04s/it]\n",
      "Loss: 0.02037:  81%|████████  | 202/250 [06:52<01:37,  2.04s/it]\n",
      "Loss: 0.01804:  81%|████████  | 202/250 [06:53<01:37,  2.04s/it]\n",
      "Loss: 0.01699:  81%|████████  | 202/250 [06:53<01:37,  2.04s/it]\n",
      "Loss: 0.02058:  81%|████████  | 202/250 [06:53<01:37,  2.04s/it]\n",
      "Loss: 0.01604:  81%|████████  | 202/250 [06:53<01:37,  2.04s/it]\n",
      "Loss: 0.01718:  81%|████████  | 202/250 [06:53<01:37,  2.04s/it]\n",
      "Loss: 0.01526:  81%|████████  | 202/250 [06:53<01:37,  2.04s/it]\n",
      "Loss: 0.01526:  81%|████████  | 203/250 [06:53<01:35,  2.04s/it]\n",
      "Loss: 0.01765:  81%|████████  | 203/250 [06:53<01:35,  2.04s/it]\n",
      "Loss: 0.01672:  81%|████████  | 203/250 [06:53<01:35,  2.04s/it]\n",
      "Loss: 0.01602:  81%|████████  | 203/250 [06:53<01:35,  2.04s/it]\n",
      "Loss: 0.01680:  81%|████████  | 203/250 [06:54<01:35,  2.04s/it]\n",
      "Loss: 0.01819:  81%|████████  | 203/250 [06:54<01:35,  2.04s/it]\n",
      "Loss: 0.01829:  81%|████████  | 203/250 [06:54<01:35,  2.04s/it]\n",
      "Loss: 0.01959:  81%|████████  | 203/250 [06:54<01:35,  2.04s/it]\n",
      "Loss: 0.01823:  81%|████████  | 203/250 [06:54<01:35,  2.04s/it]\n",
      "Loss: 0.01891:  81%|████████  | 203/250 [06:54<01:35,  2.04s/it]\n",
      "Loss: 0.01787:  81%|████████  | 203/250 [06:54<01:35,  2.04s/it]\n",
      "Loss: 0.01699:  81%|████████  | 203/250 [06:54<01:35,  2.04s/it]\n",
      "Loss: 0.01713:  81%|████████  | 203/250 [06:54<01:35,  2.04s/it]\n",
      "Loss: 0.01850:  81%|████████  | 203/250 [06:55<01:35,  2.04s/it]\n",
      "Loss: 0.01570:  81%|████████  | 203/250 [06:55<01:35,  2.04s/it]\n",
      "Loss: 0.01785:  81%|████████  | 203/250 [06:55<01:35,  2.04s/it]\n",
      "Loss: 0.01838:  81%|████████  | 203/250 [06:55<01:35,  2.04s/it]\n",
      "Loss: 0.01933:  81%|████████  | 203/250 [06:55<01:35,  2.04s/it]\n",
      "Loss: 0.01640:  81%|████████  | 203/250 [06:55<01:35,  2.04s/it]\n",
      "Loss: 0.01547:  81%|████████  | 203/250 [06:55<01:35,  2.04s/it]\n",
      "Loss: 0.01547:  82%|████████▏ | 204/250 [06:55<01:33,  2.03s/it]\n",
      "Loss: 0.01749:  82%|████████▏ | 204/250 [06:55<01:33,  2.03s/it]\n",
      "Loss: 0.01726:  82%|████████▏ | 204/250 [06:55<01:33,  2.03s/it]\n",
      "Loss: 0.01910:  82%|████████▏ | 204/250 [06:55<01:33,  2.03s/it]\n",
      "Loss: 0.01738:  82%|████████▏ | 204/250 [06:56<01:33,  2.03s/it]\n",
      "Loss: 0.01709:  82%|████████▏ | 204/250 [06:56<01:33,  2.03s/it]\n",
      "Loss: 0.01600:  82%|████████▏ | 204/250 [06:56<01:33,  2.03s/it]\n",
      "Loss: 0.01701:  82%|████████▏ | 204/250 [06:56<01:33,  2.03s/it]\n",
      "Loss: 0.01857:  82%|████████▏ | 204/250 [06:56<01:33,  2.03s/it]\n",
      "Loss: 0.01728:  82%|████████▏ | 204/250 [06:56<01:33,  2.03s/it]\n",
      "Loss: 0.01862:  82%|████████▏ | 204/250 [06:56<01:33,  2.03s/it]\n",
      "Loss: 0.01898:  82%|████████▏ | 204/250 [06:56<01:33,  2.03s/it]\n",
      "Loss: 0.01859:  82%|████████▏ | 204/250 [06:56<01:33,  2.03s/it]\n",
      "Loss: 0.01642:  82%|████████▏ | 204/250 [06:57<01:33,  2.03s/it]\n",
      "Loss: 0.01737:  82%|████████▏ | 204/250 [06:57<01:33,  2.03s/it]\n",
      "Loss: 0.01590:  82%|████████▏ | 204/250 [06:57<01:33,  2.03s/it]\n",
      "Loss: 0.01637:  82%|████████▏ | 204/250 [06:57<01:33,  2.03s/it]\n",
      "Loss: 0.01765:  82%|████████▏ | 204/250 [06:57<01:33,  2.03s/it]\n",
      "Loss: 0.01559:  82%|████████▏ | 204/250 [06:57<01:33,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 62.44it/s]\u001b[A\n",
      "Loss: 0.01559:  82%|████████▏ | 205/250 [06:57<01:31,  2.03s/it]\n",
      "Loss: 0.01740:  82%|████████▏ | 205/250 [06:57<01:31,  2.03s/it]\n",
      "Loss: 0.01723:  82%|████████▏ | 205/250 [06:57<01:31,  2.03s/it]\n",
      "Loss: 0.01997:  82%|████████▏ | 205/250 [06:58<01:31,  2.03s/it]\n",
      "Loss: 0.01676:  82%|████████▏ | 205/250 [06:58<01:31,  2.03s/it]\n",
      "Loss: 0.01708:  82%|████████▏ | 205/250 [06:58<01:31,  2.03s/it]\n",
      "Loss: 0.01517:  82%|████████▏ | 205/250 [06:58<01:31,  2.03s/it]\n",
      "Loss: 0.01599:  82%|████████▏ | 205/250 [06:58<01:31,  2.03s/it]\n",
      "Loss: 0.01730:  82%|████████▏ | 205/250 [06:58<01:31,  2.03s/it]\n",
      "Loss: 0.01833:  82%|████████▏ | 205/250 [06:58<01:31,  2.03s/it]\n",
      "Loss: 0.01822:  82%|████████▏ | 205/250 [06:58<01:31,  2.03s/it]\n",
      "Loss: 0.01666:  82%|████████▏ | 205/250 [06:58<01:31,  2.03s/it]\n",
      "Loss: 0.01740:  82%|████████▏ | 205/250 [06:58<01:31,  2.03s/it]\n",
      "Loss: 0.01852:  82%|████████▏ | 205/250 [06:59<01:31,  2.03s/it]\n",
      "Loss: 0.01555:  82%|████████▏ | 205/250 [06:59<01:31,  2.03s/it]\n",
      "Loss: 0.01786:  82%|████████▏ | 205/250 [06:59<01:31,  2.03s/it]\n",
      "Loss: 0.01861:  82%|████████▏ | 205/250 [06:59<01:31,  2.03s/it]\n",
      "Loss: 0.01915:  82%|████████▏ | 205/250 [06:59<01:31,  2.03s/it]\n",
      "Loss: 0.01614:  82%|████████▏ | 205/250 [06:59<01:31,  2.03s/it]\n",
      "Loss: 0.01536:  82%|████████▏ | 205/250 [06:59<01:31,  2.03s/it]\n",
      "Loss: 0.01536:  82%|████████▏ | 206/250 [06:59<01:29,  2.03s/it]\n",
      "Loss: 0.01745:  82%|████████▏ | 206/250 [06:59<01:29,  2.03s/it]\n",
      "Loss: 0.01718:  82%|████████▏ | 206/250 [06:59<01:29,  2.03s/it]\n",
      "Loss: 0.01937:  82%|████████▏ | 206/250 [07:00<01:29,  2.03s/it]\n",
      "Loss: 0.01678:  82%|████████▏ | 206/250 [07:00<01:29,  2.03s/it]\n",
      "Loss: 0.01496:  82%|████████▏ | 206/250 [07:00<01:29,  2.03s/it]\n",
      "Loss: 0.01835:  82%|████████▏ | 206/250 [07:00<01:29,  2.03s/it]\n",
      "Loss: 0.01962:  82%|████████▏ | 206/250 [07:00<01:29,  2.03s/it]\n",
      "Loss: 0.01808:  82%|████████▏ | 206/250 [07:00<01:29,  2.03s/it]\n",
      "Loss: 0.01834:  82%|████████▏ | 206/250 [07:00<01:29,  2.03s/it]\n",
      "Loss: 0.01830:  82%|████████▏ | 206/250 [07:00<01:29,  2.03s/it]\n",
      "Loss: 0.01723:  82%|████████▏ | 206/250 [07:00<01:29,  2.03s/it]\n",
      "Loss: 0.01755:  82%|████████▏ | 206/250 [07:01<01:29,  2.03s/it]\n",
      "Loss: 0.01848:  82%|████████▏ | 206/250 [07:01<01:29,  2.03s/it]\n",
      "Loss: 0.01541:  82%|████████▏ | 206/250 [07:01<01:29,  2.03s/it]\n",
      "Loss: 0.01789:  82%|████████▏ | 206/250 [07:01<01:29,  2.03s/it]\n",
      "Loss: 0.01878:  82%|████████▏ | 206/250 [07:01<01:29,  2.03s/it]\n",
      "Loss: 0.01959:  82%|████████▏ | 206/250 [07:01<01:29,  2.03s/it]\n",
      "Loss: 0.01628:  82%|████████▏ | 206/250 [07:01<01:29,  2.03s/it]\n",
      "Loss: 0.01520:  82%|████████▏ | 206/250 [07:01<01:29,  2.03s/it]\n",
      "Loss: 0.01520:  83%|████████▎ | 207/250 [07:01<01:27,  2.04s/it]\n",
      "Loss: 0.01753:  83%|████████▎ | 207/250 [07:01<01:27,  2.04s/it]\n",
      "Loss: 0.01731:  83%|████████▎ | 207/250 [07:01<01:27,  2.04s/it]\n",
      "Loss: 0.01889:  83%|████████▎ | 207/250 [07:02<01:27,  2.04s/it]\n",
      "Loss: 0.01631:  83%|████████▎ | 207/250 [07:02<01:27,  2.04s/it]\n",
      "Loss: 0.01698:  83%|████████▎ | 207/250 [07:02<01:27,  2.04s/it]\n",
      "Loss: 0.01795:  83%|████████▎ | 207/250 [07:02<01:27,  2.04s/it]\n",
      "Loss: 0.01599:  83%|████████▎ | 207/250 [07:02<01:27,  2.04s/it]\n",
      "Loss: 0.01795:  83%|████████▎ | 207/250 [07:02<01:27,  2.04s/it]\n",
      "Loss: 0.01800:  83%|████████▎ | 207/250 [07:02<01:27,  2.04s/it]\n",
      "Loss: 0.01732:  83%|████████▎ | 207/250 [07:02<01:27,  2.04s/it]\n",
      "Loss: 0.01983:  83%|████████▎ | 207/250 [07:02<01:27,  2.04s/it]\n",
      "Loss: 0.01717:  83%|████████▎ | 207/250 [07:03<01:27,  2.04s/it]\n",
      "Loss: 0.01841:  83%|████████▎ | 207/250 [07:03<01:27,  2.04s/it]\n",
      "Loss: 0.01926:  83%|████████▎ | 207/250 [07:03<01:27,  2.04s/it]\n",
      "Loss: 0.01776:  83%|████████▎ | 207/250 [07:03<01:27,  2.04s/it]\n",
      "Loss: 0.01592:  83%|████████▎ | 207/250 [07:03<01:27,  2.04s/it]\n",
      "Loss: 0.01864:  83%|████████▎ | 207/250 [07:03<01:27,  2.04s/it]\n",
      "Loss: 0.01898:  83%|████████▎ | 207/250 [07:03<01:27,  2.04s/it]\n",
      "Loss: 0.01519:  83%|████████▎ | 207/250 [07:03<01:27,  2.04s/it]\n",
      "Loss: 0.01519:  83%|████████▎ | 208/250 [07:03<01:25,  2.05s/it]\n",
      "Loss: 0.01739:  83%|████████▎ | 208/250 [07:03<01:25,  2.05s/it]\n",
      "Loss: 0.01732:  83%|████████▎ | 208/250 [07:04<01:25,  2.05s/it]\n",
      "Loss: 0.01874:  83%|████████▎ | 208/250 [07:04<01:25,  2.05s/it]\n",
      "Loss: 0.01754:  83%|████████▎ | 208/250 [07:04<01:25,  2.05s/it]\n",
      "Loss: 0.01687:  83%|████████▎ | 208/250 [07:04<01:25,  2.05s/it]\n",
      "Loss: 0.01789:  83%|████████▎ | 208/250 [07:04<01:25,  2.05s/it]\n",
      "Loss: 0.01945:  83%|████████▎ | 208/250 [07:04<01:25,  2.05s/it]\n",
      "Loss: 0.01907:  83%|████████▎ | 208/250 [07:04<01:25,  2.05s/it]\n",
      "Loss: 0.01864:  83%|████████▎ | 208/250 [07:04<01:25,  2.05s/it]\n",
      "Loss: 0.01850:  83%|████████▎ | 208/250 [07:04<01:25,  2.05s/it]\n",
      "Loss: 0.01693:  83%|████████▎ | 208/250 [07:05<01:25,  2.05s/it]\n",
      "Loss: 0.01612:  83%|████████▎ | 208/250 [07:05<01:25,  2.05s/it]\n",
      "Loss: 0.01773:  83%|████████▎ | 208/250 [07:05<01:25,  2.05s/it]\n",
      "Loss: 0.01641:  83%|████████▎ | 208/250 [07:05<01:25,  2.05s/it]\n",
      "Loss: 0.01687:  83%|████████▎ | 208/250 [07:05<01:25,  2.05s/it]\n",
      "Loss: 0.01726:  83%|████████▎ | 208/250 [07:05<01:25,  2.05s/it]\n",
      "Loss: 0.01894:  83%|████████▎ | 208/250 [07:05<01:25,  2.05s/it]\n",
      "Loss: 0.01619:  83%|████████▎ | 208/250 [07:05<01:25,  2.05s/it]\n",
      "Loss: 0.01529:  83%|████████▎ | 208/250 [07:05<01:25,  2.05s/it]\n",
      "Loss: 0.01529:  84%|████████▎ | 209/250 [07:05<01:24,  2.05s/it]\n",
      "Loss: 0.01674:  84%|████████▎ | 209/250 [07:05<01:24,  2.05s/it]\n",
      "Loss: 0.01635:  84%|████████▎ | 209/250 [07:06<01:24,  2.05s/it]\n",
      "Loss: 0.01624:  84%|████████▎ | 209/250 [07:06<01:24,  2.05s/it]\n",
      "Loss: 0.01910:  84%|████████▎ | 209/250 [07:06<01:24,  2.05s/it]\n",
      "Loss: 0.01552:  84%|████████▎ | 209/250 [07:06<01:24,  2.05s/it]\n",
      "Loss: 0.01817:  84%|████████▎ | 209/250 [07:06<01:24,  2.05s/it]\n",
      "Loss: 0.01956:  84%|████████▎ | 209/250 [07:06<01:24,  2.05s/it]\n",
      "Loss: 0.01889:  84%|████████▎ | 209/250 [07:06<01:24,  2.05s/it]\n",
      "Loss: 0.01689:  84%|████████▎ | 209/250 [07:06<01:24,  2.05s/it]\n",
      "Loss: 0.01874:  84%|████████▎ | 209/250 [07:06<01:24,  2.05s/it]\n",
      "Loss: 0.01682:  84%|████████▎ | 209/250 [07:07<01:24,  2.05s/it]\n",
      "Loss: 0.01916:  84%|████████▎ | 209/250 [07:07<01:24,  2.05s/it]\n",
      "Loss: 0.01654:  84%|████████▎ | 209/250 [07:07<01:24,  2.05s/it]\n",
      "Loss: 0.01699:  84%|████████▎ | 209/250 [07:07<01:24,  2.05s/it]\n",
      "Loss: 0.01722:  84%|████████▎ | 209/250 [07:07<01:24,  2.05s/it]\n",
      "Loss: 0.01622:  84%|████████▎ | 209/250 [07:07<01:24,  2.05s/it]\n",
      "Loss: 0.01832:  84%|████████▎ | 209/250 [07:07<01:24,  2.05s/it]\n",
      "Loss: 0.01766:  84%|████████▎ | 209/250 [07:07<01:24,  2.05s/it]\n",
      "Loss: 0.01514:  84%|████████▎ | 209/250 [07:07<01:24,  2.05s/it]\n",
      "Loss: 0.01514:  84%|████████▍ | 210/250 [07:07<01:22,  2.06s/it]\n",
      "Loss: 0.01767:  84%|████████▍ | 210/250 [07:08<01:22,  2.06s/it]\n",
      "Loss: 0.01779:  84%|████████▍ | 210/250 [07:08<01:22,  2.06s/it]\n",
      "Loss: 0.01867:  84%|████████▍ | 210/250 [07:08<01:22,  2.06s/it]\n",
      "Loss: 0.01637:  84%|████████▍ | 210/250 [07:08<01:22,  2.06s/it]\n",
      "Loss: 0.01721:  84%|████████▍ | 210/250 [07:08<01:22,  2.06s/it]\n",
      "Loss: 0.01531:  84%|████████▍ | 210/250 [07:08<01:22,  2.06s/it]\n",
      "Loss: 0.01687:  84%|████████▍ | 210/250 [07:08<01:22,  2.06s/it]\n",
      "Loss: 0.01742:  84%|████████▍ | 210/250 [07:08<01:22,  2.06s/it]\n",
      "Loss: 0.01859:  84%|████████▍ | 210/250 [07:08<01:22,  2.06s/it]\n",
      "Loss: 0.01710:  84%|████████▍ | 210/250 [07:09<01:22,  2.06s/it]\n",
      "Loss: 0.01616:  84%|████████▍ | 210/250 [07:09<01:22,  2.06s/it]\n",
      "Loss: 0.01929:  84%|████████▍ | 210/250 [07:09<01:22,  2.06s/it]\n",
      "Loss: 0.01710:  84%|████████▍ | 210/250 [07:09<01:22,  2.06s/it]\n",
      "Loss: 0.01833:  84%|████████▍ | 210/250 [07:09<01:22,  2.06s/it]\n",
      "Loss: 0.01571:  84%|████████▍ | 210/250 [07:09<01:22,  2.06s/it]\n",
      "Loss: 0.01698:  84%|████████▍ | 210/250 [07:09<01:22,  2.06s/it]\n",
      "Loss: 0.01494:  84%|████████▍ | 210/250 [07:09<01:22,  2.06s/it]\n",
      "Loss: 0.01784:  84%|████████▍ | 210/250 [07:09<01:22,  2.06s/it]\n",
      "Loss: 0.01544:  84%|████████▍ | 210/250 [07:10<01:22,  2.06s/it]\n",
      "Loss: 0.01544:  84%|████████▍ | 211/250 [07:10<01:20,  2.06s/it]\n",
      "Loss: 0.01653:  84%|████████▍ | 211/250 [07:10<01:20,  2.06s/it]\n",
      "Loss: 0.01610:  84%|████████▍ | 211/250 [07:10<01:20,  2.06s/it]\n",
      "Loss: 0.02002:  84%|████████▍ | 211/250 [07:10<01:20,  2.06s/it]\n",
      "Loss: 0.01601:  84%|████████▍ | 211/250 [07:10<01:20,  2.06s/it]\n",
      "Loss: 0.01711:  84%|████████▍ | 211/250 [07:10<01:20,  2.06s/it]\n",
      "Loss: 0.01549:  84%|████████▍ | 211/250 [07:10<01:20,  2.06s/it]\n",
      "Loss: 0.01693:  84%|████████▍ | 211/250 [07:10<01:20,  2.06s/it]\n",
      "Loss: 0.01854:  84%|████████▍ | 211/250 [07:10<01:20,  2.06s/it]\n",
      "Loss: 0.01844:  84%|████████▍ | 211/250 [07:11<01:20,  2.06s/it]\n",
      "Loss: 0.01838:  84%|████████▍ | 211/250 [07:11<01:20,  2.06s/it]\n",
      "Loss: 0.01848:  84%|████████▍ | 211/250 [07:11<01:20,  2.06s/it]\n",
      "Loss: 0.01852:  84%|████████▍ | 211/250 [07:11<01:20,  2.06s/it]\n",
      "Loss: 0.01621:  84%|████████▍ | 211/250 [07:11<01:20,  2.06s/it]\n",
      "Loss: 0.01709:  84%|████████▍ | 211/250 [07:11<01:20,  2.06s/it]\n",
      "Loss: 0.01605:  84%|████████▍ | 211/250 [07:11<01:20,  2.06s/it]\n",
      "Loss: 0.01610:  84%|████████▍ | 211/250 [07:11<01:20,  2.06s/it]\n",
      "Loss: 0.01746:  84%|████████▍ | 211/250 [07:11<01:20,  2.06s/it]\n",
      "Loss: 0.01496:  84%|████████▍ | 211/250 [07:12<01:20,  2.06s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 62.77it/s]\u001b[A\n",
      "Loss: 0.01496:  85%|████████▍ | 212/250 [07:12<01:18,  2.06s/it]\n",
      "Loss: 0.01753:  85%|████████▍ | 212/250 [07:12<01:18,  2.06s/it]\n",
      "Loss: 0.01732:  85%|████████▍ | 212/250 [07:12<01:18,  2.06s/it]\n",
      "Loss: 0.01871:  85%|████████▍ | 212/250 [07:12<01:18,  2.06s/it]\n",
      "Loss: 0.01753:  85%|████████▍ | 212/250 [07:12<01:18,  2.06s/it]\n",
      "Loss: 0.01703:  85%|████████▍ | 212/250 [07:12<01:18,  2.06s/it]\n",
      "Loss: 0.01755:  85%|████████▍ | 212/250 [07:12<01:18,  2.06s/it]\n",
      "Loss: 0.01765:  85%|████████▍ | 212/250 [07:12<01:18,  2.06s/it]\n",
      "Loss: 0.01854:  85%|████████▍ | 212/250 [07:12<01:18,  2.06s/it]\n",
      "Loss: 0.01667:  85%|████████▍ | 212/250 [07:13<01:18,  2.06s/it]\n",
      "Loss: 0.01836:  85%|████████▍ | 212/250 [07:13<01:18,  2.06s/it]\n",
      "Loss: 0.01925:  85%|████████▍ | 212/250 [07:13<01:18,  2.06s/it]\n",
      "Loss: 0.01655:  85%|████████▍ | 212/250 [07:13<01:18,  2.06s/it]\n",
      "Loss: 0.01715:  85%|████████▍ | 212/250 [07:13<01:18,  2.06s/it]\n",
      "Loss: 0.01728:  85%|████████▍ | 212/250 [07:13<01:18,  2.06s/it]\n",
      "Loss: 0.01651:  85%|████████▍ | 212/250 [07:13<01:18,  2.06s/it]\n",
      "Loss: 0.01864:  85%|████████▍ | 212/250 [07:13<01:18,  2.06s/it]\n",
      "Loss: 0.01784:  85%|████████▍ | 212/250 [07:13<01:18,  2.06s/it]\n",
      "Loss: 0.01518:  85%|████████▍ | 212/250 [07:14<01:18,  2.06s/it]\n",
      "Loss: 0.01518:  85%|████████▌ | 213/250 [07:14<01:15,  2.05s/it]\n",
      "Loss: 0.01722:  85%|████████▌ | 213/250 [07:14<01:15,  2.05s/it]\n",
      "Loss: 0.01737:  85%|████████▌ | 213/250 [07:14<01:15,  2.05s/it]\n",
      "Loss: 0.01924:  85%|████████▌ | 213/250 [07:14<01:15,  2.05s/it]\n",
      "Loss: 0.01628:  85%|████████▌ | 213/250 [07:14<01:15,  2.05s/it]\n",
      "Loss: 0.01737:  85%|████████▌ | 213/250 [07:14<01:15,  2.05s/it]\n",
      "Loss: 0.01504:  85%|████████▌ | 213/250 [07:14<01:15,  2.05s/it]\n",
      "Loss: 0.01624:  85%|████████▌ | 213/250 [07:14<01:15,  2.05s/it]\n",
      "Loss: 0.01702:  85%|████████▌ | 213/250 [07:14<01:15,  2.05s/it]\n",
      "Loss: 0.01828:  85%|████████▌ | 213/250 [07:15<01:15,  2.05s/it]\n",
      "Loss: 0.01766:  85%|████████▌ | 213/250 [07:15<01:15,  2.05s/it]\n",
      "Loss: 0.01667:  85%|████████▌ | 213/250 [07:15<01:15,  2.05s/it]\n",
      "Loss: 0.01927:  85%|████████▌ | 213/250 [07:15<01:15,  2.05s/it]\n",
      "Loss: 0.01993:  85%|████████▌ | 213/250 [07:15<01:15,  2.05s/it]\n",
      "Loss: 0.01817:  85%|████████▌ | 213/250 [07:15<01:15,  2.05s/it]\n",
      "Loss: 0.01731:  85%|████████▌ | 213/250 [07:15<01:15,  2.05s/it]\n",
      "Loss: 0.02087:  85%|████████▌ | 213/250 [07:15<01:15,  2.05s/it]\n",
      "Loss: 0.01563:  85%|████████▌ | 213/250 [07:15<01:15,  2.05s/it]\n",
      "Loss: 0.01716:  85%|████████▌ | 213/250 [07:16<01:15,  2.05s/it]\n",
      "Loss: 0.01499:  85%|████████▌ | 213/250 [07:16<01:15,  2.05s/it]\n",
      "Loss: 0.01499:  86%|████████▌ | 214/250 [07:16<01:13,  2.05s/it]\n",
      "Loss: 0.01715:  86%|████████▌ | 214/250 [07:16<01:13,  2.05s/it]\n",
      "Loss: 0.01680:  86%|████████▌ | 214/250 [07:16<01:13,  2.05s/it]\n",
      "Loss: 0.01940:  86%|████████▌ | 214/250 [07:16<01:13,  2.05s/it]\n",
      "Loss: 0.01627:  86%|████████▌ | 214/250 [07:16<01:13,  2.05s/it]\n",
      "Loss: 0.01732:  86%|████████▌ | 214/250 [07:16<01:13,  2.05s/it]\n",
      "Loss: 0.01531:  86%|████████▌ | 214/250 [07:16<01:13,  2.05s/it]\n",
      "Loss: 0.01682:  86%|████████▌ | 214/250 [07:16<01:13,  2.05s/it]\n",
      "Loss: 0.01854:  86%|████████▌ | 214/250 [07:17<01:13,  2.05s/it]\n",
      "Loss: 0.01780:  86%|████████▌ | 214/250 [07:17<01:13,  2.05s/it]\n",
      "Loss: 0.01788:  86%|████████▌ | 214/250 [07:17<01:13,  2.05s/it]\n",
      "Loss: 0.01721:  86%|████████▌ | 214/250 [07:17<01:13,  2.05s/it]\n",
      "Loss: 0.01719:  86%|████████▌ | 214/250 [07:17<01:13,  2.05s/it]\n",
      "Loss: 0.01669:  86%|████████▌ | 214/250 [07:17<01:13,  2.05s/it]\n",
      "Loss: 0.01895:  86%|████████▌ | 214/250 [07:17<01:13,  2.05s/it]\n",
      "Loss: 0.01566:  86%|████████▌ | 214/250 [07:17<01:13,  2.05s/it]\n",
      "Loss: 0.01701:  86%|████████▌ | 214/250 [07:17<01:13,  2.05s/it]\n",
      "Loss: 0.01483:  86%|████████▌ | 214/250 [07:18<01:13,  2.05s/it]\n",
      "Loss: 0.01767:  86%|████████▌ | 214/250 [07:18<01:13,  2.05s/it]\n",
      "Loss: 0.01500:  86%|████████▌ | 214/250 [07:18<01:13,  2.05s/it]\n",
      "Loss: 0.01500:  86%|████████▌ | 215/250 [07:18<01:11,  2.05s/it]\n",
      "Loss: 0.01774:  86%|████████▌ | 215/250 [07:18<01:11,  2.05s/it]\n",
      "Loss: 0.01722:  86%|████████▌ | 215/250 [07:18<01:11,  2.05s/it]\n",
      "Loss: 0.01922:  86%|████████▌ | 215/250 [07:18<01:11,  2.05s/it]\n",
      "Loss: 0.01638:  86%|████████▌ | 215/250 [07:18<01:11,  2.05s/it]\n",
      "Loss: 0.01699:  86%|████████▌ | 215/250 [07:18<01:11,  2.05s/it]\n",
      "Loss: 0.01510:  86%|████████▌ | 215/250 [07:18<01:11,  2.05s/it]\n",
      "Loss: 0.01581:  86%|████████▌ | 215/250 [07:19<01:11,  2.05s/it]\n",
      "Loss: 0.01772:  86%|████████▌ | 215/250 [07:19<01:11,  2.05s/it]\n",
      "Loss: 0.01820:  86%|████████▌ | 215/250 [07:19<01:11,  2.05s/it]\n",
      "Loss: 0.01747:  86%|████████▌ | 215/250 [07:19<01:11,  2.05s/it]\n",
      "Loss: 0.01680:  86%|████████▌ | 215/250 [07:19<01:11,  2.05s/it]\n",
      "Loss: 0.01701:  86%|████████▌ | 215/250 [07:19<01:11,  2.05s/it]\n",
      "Loss: 0.01815:  86%|████████▌ | 215/250 [07:19<01:11,  2.05s/it]\n",
      "Loss: 0.01545:  86%|████████▌ | 215/250 [07:19<01:11,  2.05s/it]\n",
      "Loss: 0.01764:  86%|████████▌ | 215/250 [07:19<01:11,  2.05s/it]\n",
      "Loss: 0.01822:  86%|████████▌ | 215/250 [07:20<01:11,  2.05s/it]\n",
      "Loss: 0.01921:  86%|████████▌ | 215/250 [07:20<01:11,  2.05s/it]\n",
      "Loss: 0.01605:  86%|████████▌ | 215/250 [07:20<01:11,  2.05s/it]\n",
      "Loss: 0.01521:  86%|████████▌ | 215/250 [07:20<01:11,  2.05s/it]\n",
      "Loss: 0.01521:  86%|████████▋ | 216/250 [07:20<01:09,  2.06s/it]\n",
      "Loss: 0.01693:  86%|████████▋ | 216/250 [07:20<01:09,  2.06s/it]\n",
      "Loss: 0.01716:  86%|████████▋ | 216/250 [07:20<01:09,  2.06s/it]\n",
      "Loss: 0.01892:  86%|████████▋ | 216/250 [07:20<01:09,  2.06s/it]\n",
      "Loss: 0.01757:  86%|████████▋ | 216/250 [07:20<01:09,  2.06s/it]\n",
      "Loss: 0.01681:  86%|████████▋ | 216/250 [07:20<01:09,  2.06s/it]\n",
      "Loss: 0.01722:  86%|████████▋ | 216/250 [07:20<01:09,  2.06s/it]\n",
      "Loss: 0.01795:  86%|████████▋ | 216/250 [07:21<01:09,  2.06s/it]\n",
      "Loss: 0.01779:  86%|████████▋ | 216/250 [07:21<01:09,  2.06s/it]\n",
      "Loss: 0.01606:  86%|████████▋ | 216/250 [07:21<01:09,  2.06s/it]\n",
      "Loss: 0.01840:  86%|████████▋ | 216/250 [07:21<01:09,  2.06s/it]\n",
      "Loss: 0.01934:  86%|████████▋ | 216/250 [07:21<01:09,  2.06s/it]\n",
      "Loss: 0.01773:  86%|████████▋ | 216/250 [07:21<01:09,  2.06s/it]\n",
      "Loss: 0.01754:  86%|████████▋ | 216/250 [07:21<01:09,  2.06s/it]\n",
      "Loss: 0.01734:  86%|████████▋ | 216/250 [07:21<01:09,  2.06s/it]\n",
      "Loss: 0.01674:  86%|████████▋ | 216/250 [07:21<01:09,  2.06s/it]\n",
      "Loss: 0.01822:  86%|████████▋ | 216/250 [07:22<01:09,  2.06s/it]\n",
      "Loss: 0.01731:  86%|████████▋ | 216/250 [07:22<01:09,  2.06s/it]\n",
      "Loss: 0.01492:  86%|████████▋ | 216/250 [07:22<01:09,  2.06s/it]\n",
      "Loss: 0.01492:  87%|████████▋ | 217/250 [07:22<01:07,  2.05s/it]\n",
      "Loss: 0.01762:  87%|████████▋ | 217/250 [07:22<01:07,  2.05s/it]\n",
      "Loss: 0.01731:  87%|████████▋ | 217/250 [07:22<01:07,  2.05s/it]\n",
      "Loss: 0.01917:  87%|████████▋ | 217/250 [07:22<01:07,  2.05s/it]\n",
      "Loss: 0.01647:  87%|████████▋ | 217/250 [07:22<01:07,  2.05s/it]\n",
      "Loss: 0.01536:  87%|████████▋ | 217/250 [07:22<01:07,  2.05s/it]\n",
      "Loss: 0.01811:  87%|████████▋ | 217/250 [07:22<01:07,  2.05s/it]\n",
      "Loss: 0.01582:  87%|████████▋ | 217/250 [07:23<01:07,  2.05s/it]\n",
      "Loss: 0.01686:  87%|████████▋ | 217/250 [07:23<01:07,  2.05s/it]\n",
      "Loss: 0.01744:  87%|████████▋ | 217/250 [07:23<01:07,  2.05s/it]\n",
      "Loss: 0.01736:  87%|████████▋ | 217/250 [07:23<01:07,  2.05s/it]\n",
      "Loss: 0.01963:  87%|████████▋ | 217/250 [07:23<01:07,  2.05s/it]\n",
      "Loss: 0.01729:  87%|████████▋ | 217/250 [07:23<01:07,  2.05s/it]\n",
      "Loss: 0.01797:  87%|████████▋ | 217/250 [07:23<01:07,  2.05s/it]\n",
      "Loss: 0.01905:  87%|████████▋ | 217/250 [07:23<01:07,  2.05s/it]\n",
      "Loss: 0.01756:  87%|████████▋ | 217/250 [07:23<01:07,  2.05s/it]\n",
      "Loss: 0.01648:  87%|████████▋ | 217/250 [07:24<01:07,  2.05s/it]\n",
      "Loss: 0.01852:  87%|████████▋ | 217/250 [07:24<01:07,  2.05s/it]\n",
      "Loss: 0.01899:  87%|████████▋ | 217/250 [07:24<01:07,  2.05s/it]\n",
      "Loss: 0.01496:  87%|████████▋ | 217/250 [07:24<01:07,  2.05s/it]\n",
      "Loss: 0.01496:  87%|████████▋ | 218/250 [07:24<01:05,  2.04s/it]\n",
      "Loss: 0.01745:  87%|████████▋ | 218/250 [07:24<01:05,  2.04s/it]\n",
      "Loss: 0.01722:  87%|████████▋ | 218/250 [07:24<01:05,  2.04s/it]\n",
      "Loss: 0.01938:  87%|████████▋ | 218/250 [07:24<01:05,  2.04s/it]\n",
      "Loss: 0.01860:  87%|████████▋ | 218/250 [07:24<01:05,  2.04s/it]\n",
      "Loss: 0.01553:  87%|████████▋ | 218/250 [07:24<01:05,  2.04s/it]\n",
      "Loss: 0.01785:  87%|████████▋ | 218/250 [07:25<01:05,  2.04s/it]\n",
      "Loss: 0.01940:  87%|████████▋ | 218/250 [07:25<01:05,  2.04s/it]\n",
      "Loss: 0.01874:  87%|████████▋ | 218/250 [07:25<01:05,  2.04s/it]\n",
      "Loss: 0.01867:  87%|████████▋ | 218/250 [07:25<01:05,  2.04s/it]\n",
      "Loss: 0.01711:  87%|████████▋ | 218/250 [07:25<01:05,  2.04s/it]\n",
      "Loss: 0.01646:  87%|████████▋ | 218/250 [07:25<01:05,  2.04s/it]\n",
      "Loss: 0.01929:  87%|████████▋ | 218/250 [07:25<01:05,  2.04s/it]\n",
      "Loss: 0.01629:  87%|████████▋ | 218/250 [07:25<01:05,  2.04s/it]\n",
      "Loss: 0.01851:  87%|████████▋ | 218/250 [07:25<01:05,  2.04s/it]\n",
      "Loss: 0.01589:  87%|████████▋ | 218/250 [07:26<01:05,  2.04s/it]\n",
      "Loss: 0.02083:  87%|████████▋ | 218/250 [07:26<01:05,  2.04s/it]\n",
      "Loss: 0.01540:  87%|████████▋ | 218/250 [07:26<01:05,  2.04s/it]\n",
      "Loss: 0.01780:  87%|████████▋ | 218/250 [07:26<01:05,  2.04s/it]\n",
      "Loss: 0.01517:  87%|████████▋ | 218/250 [07:26<01:05,  2.04s/it]\n",
      "Loss: 0.01517:  88%|████████▊ | 219/250 [07:26<01:03,  2.06s/it]\n",
      "Loss: 0.01733:  88%|████████▊ | 219/250 [07:26<01:03,  2.06s/it]\n",
      "Loss: 0.01764:  88%|████████▊ | 219/250 [07:26<01:03,  2.06s/it]\n",
      "Loss: 0.01879:  88%|████████▊ | 219/250 [07:26<01:03,  2.06s/it]\n",
      "Loss: 0.01729:  88%|████████▊ | 219/250 [07:26<01:03,  2.06s/it]\n",
      "Loss: 0.01711:  88%|████████▊ | 219/250 [07:27<01:03,  2.06s/it]\n",
      "Loss: 0.01698:  88%|████████▊ | 219/250 [07:27<01:03,  2.06s/it]\n",
      "Loss: 0.01696:  88%|████████▊ | 219/250 [07:27<01:03,  2.06s/it]\n",
      "Loss: 0.01852:  88%|████████▊ | 219/250 [07:27<01:03,  2.06s/it]\n",
      "Loss: 0.01751:  88%|████████▊ | 219/250 [07:27<01:03,  2.06s/it]\n",
      "Loss: 0.01720:  88%|████████▊ | 219/250 [07:27<01:03,  2.06s/it]\n",
      "Loss: 0.01959:  88%|████████▊ | 219/250 [07:27<01:03,  2.06s/it]\n",
      "Loss: 0.01727:  88%|████████▊ | 219/250 [07:27<01:03,  2.06s/it]\n",
      "Loss: 0.01759:  88%|████████▊ | 219/250 [07:27<01:03,  2.06s/it]\n",
      "Loss: 0.01867:  88%|████████▊ | 219/250 [07:27<01:03,  2.06s/it]\n",
      "Loss: 0.01782:  88%|████████▊ | 219/250 [07:28<01:03,  2.06s/it]\n",
      "Loss: 0.01641:  88%|████████▊ | 219/250 [07:28<01:03,  2.06s/it]\n",
      "Loss: 0.01879:  88%|████████▊ | 219/250 [07:28<01:03,  2.06s/it]\n",
      "Loss: 0.01897:  88%|████████▊ | 219/250 [07:28<01:03,  2.06s/it]\n",
      "Loss: 0.01502:  88%|████████▊ | 219/250 [07:28<01:03,  2.06s/it]\n",
      "Loss: 0.01502:  88%|████████▊ | 220/250 [07:28<01:01,  2.04s/it]\n",
      "Loss: 0.01749:  88%|████████▊ | 220/250 [07:28<01:01,  2.04s/it]\n",
      "Loss: 0.01607:  88%|████████▊ | 220/250 [07:28<01:01,  2.04s/it]\n",
      "Loss: 0.01932:  88%|████████▊ | 220/250 [07:28<01:01,  2.04s/it]\n",
      "Loss: 0.01646:  88%|████████▊ | 220/250 [07:28<01:01,  2.04s/it]\n",
      "Loss: 0.01482:  88%|████████▊ | 220/250 [07:28<01:01,  2.04s/it]\n",
      "Loss: 0.01795:  88%|████████▊ | 220/250 [07:29<01:01,  2.04s/it]\n",
      "Loss: 0.01570:  88%|████████▊ | 220/250 [07:29<01:01,  2.04s/it]\n",
      "Loss: 0.01759:  88%|████████▊ | 220/250 [07:29<01:01,  2.04s/it]\n",
      "Loss: 0.01871:  88%|████████▊ | 220/250 [07:29<01:01,  2.04s/it]\n",
      "Loss: 0.01770:  88%|████████▊ | 220/250 [07:29<01:01,  2.04s/it]\n",
      "Loss: 0.01690:  88%|████████▊ | 220/250 [07:29<01:01,  2.04s/it]\n",
      "Loss: 0.01703:  88%|████████▊ | 220/250 [07:29<01:01,  2.04s/it]\n",
      "Loss: 0.01814:  88%|████████▊ | 220/250 [07:29<01:01,  2.04s/it]\n",
      "Loss: 0.01585:  88%|████████▊ | 220/250 [07:29<01:01,  2.04s/it]\n",
      "Loss: 0.01772:  88%|████████▊ | 220/250 [07:30<01:01,  2.04s/it]\n",
      "Loss: 0.01826:  88%|████████▊ | 220/250 [07:30<01:01,  2.04s/it]\n",
      "Loss: 0.01956:  88%|████████▊ | 220/250 [07:30<01:01,  2.04s/it]\n",
      "Loss: 0.01608:  88%|████████▊ | 220/250 [07:30<01:01,  2.04s/it]\n",
      "Loss: 0.01537:  88%|████████▊ | 220/250 [07:30<01:01,  2.04s/it]\n",
      "Loss: 0.01537:  88%|████████▊ | 221/250 [07:30<00:59,  2.04s/it]\n",
      "Loss: 0.01718:  88%|████████▊ | 221/250 [07:30<00:59,  2.04s/it]\n",
      "Loss: 0.01621:  88%|████████▊ | 221/250 [07:30<00:59,  2.04s/it]\n",
      "Loss: 0.01927:  88%|████████▊ | 221/250 [07:30<00:59,  2.04s/it]\n",
      "Loss: 0.01869:  88%|████████▊ | 221/250 [07:30<00:59,  2.04s/it]\n",
      "Loss: 0.01494:  88%|████████▊ | 221/250 [07:31<00:59,  2.04s/it]\n",
      "Loss: 0.01777:  88%|████████▊ | 221/250 [07:31<00:59,  2.04s/it]\n",
      "Loss: 0.01611:  88%|████████▊ | 221/250 [07:31<00:59,  2.04s/it]\n",
      "Loss: 0.01697:  88%|████████▊ | 221/250 [07:31<00:59,  2.04s/it]\n",
      "Loss: 0.01783:  88%|████████▊ | 221/250 [07:31<00:59,  2.04s/it]\n",
      "Loss: 0.01725:  88%|████████▊ | 221/250 [07:31<00:59,  2.04s/it]\n",
      "Loss: 0.01952:  88%|████████▊ | 221/250 [07:31<00:59,  2.04s/it]\n",
      "Loss: 0.01723:  88%|████████▊ | 221/250 [07:31<00:59,  2.04s/it]\n",
      "Loss: 0.01756:  88%|████████▊ | 221/250 [07:31<00:59,  2.04s/it]\n",
      "Loss: 0.01858:  88%|████████▊ | 221/250 [07:32<00:59,  2.04s/it]\n",
      "Loss: 0.01736:  88%|████████▊ | 221/250 [07:32<00:59,  2.04s/it]\n",
      "Loss: 0.01588:  88%|████████▊ | 221/250 [07:32<00:59,  2.04s/it]\n",
      "Loss: 0.01816:  88%|████████▊ | 221/250 [07:32<00:59,  2.04s/it]\n",
      "Loss: 0.01854:  88%|████████▊ | 221/250 [07:32<00:59,  2.04s/it]\n",
      "Loss: 0.01535:  88%|████████▊ | 221/250 [07:32<00:59,  2.04s/it]\n",
      "Loss: 0.01535:  89%|████████▉ | 222/250 [07:32<00:56,  2.03s/it]\n",
      "Loss: 0.01739:  89%|████████▉ | 222/250 [07:32<00:56,  2.03s/it]\n",
      "Loss: 0.01615:  89%|████████▉ | 222/250 [07:32<00:56,  2.03s/it]\n",
      "Loss: 0.01586:  89%|████████▉ | 222/250 [07:32<00:56,  2.03s/it]\n",
      "Loss: 0.01890:  89%|████████▉ | 222/250 [07:32<00:56,  2.03s/it]\n",
      "Loss: 0.01546:  89%|████████▉ | 222/250 [07:33<00:56,  2.03s/it]\n",
      "Loss: 0.01782:  89%|████████▉ | 222/250 [07:33<00:56,  2.03s/it]\n",
      "Loss: 0.01564:  89%|████████▉ | 222/250 [07:33<00:56,  2.03s/it]\n",
      "Loss: 0.01738:  89%|████████▉ | 222/250 [07:33<00:56,  2.03s/it]\n",
      "Loss: 0.01791:  89%|████████▉ | 222/250 [07:33<00:56,  2.03s/it]\n",
      "Loss: 0.01749:  89%|████████▉ | 222/250 [07:33<00:56,  2.03s/it]\n",
      "Loss: 0.01646:  89%|████████▉ | 222/250 [07:33<00:56,  2.03s/it]\n",
      "Loss: 0.01717:  89%|████████▉ | 222/250 [07:33<00:56,  2.03s/it]\n",
      "Loss: 0.01737:  89%|████████▉ | 222/250 [07:33<00:56,  2.03s/it]\n",
      "Loss: 0.01534:  89%|████████▉ | 222/250 [07:34<00:56,  2.03s/it]\n",
      "Loss: 0.01770:  89%|████████▉ | 222/250 [07:34<00:56,  2.03s/it]\n",
      "Loss: 0.01848:  89%|████████▉ | 222/250 [07:34<00:56,  2.03s/it]\n",
      "Loss: 0.01919:  89%|████████▉ | 222/250 [07:34<00:56,  2.03s/it]\n",
      "Loss: 0.01593:  89%|████████▉ | 222/250 [07:34<00:56,  2.03s/it]\n",
      "Loss: 0.01512:  89%|████████▉ | 222/250 [07:34<00:56,  2.03s/it]\n",
      "Loss: 0.01512:  89%|████████▉ | 223/250 [07:34<00:54,  2.04s/it]\n",
      "Loss: 0.01719:  89%|████████▉ | 223/250 [07:34<00:54,  2.04s/it]\n",
      "Loss: 0.01713:  89%|████████▉ | 223/250 [07:34<00:54,  2.04s/it]\n",
      "Loss: 0.01891:  89%|████████▉ | 223/250 [07:34<00:54,  2.04s/it]\n",
      "Loss: 0.01666:  89%|████████▉ | 223/250 [07:34<00:54,  2.04s/it]\n",
      "Loss: 0.01711:  89%|████████▉ | 223/250 [07:35<00:54,  2.04s/it]\n",
      "Loss: 0.01530:  89%|████████▉ | 223/250 [07:35<00:54,  2.04s/it]\n",
      "Loss: 0.01660:  89%|████████▉ | 223/250 [07:35<00:54,  2.04s/it]\n",
      "Loss: 0.01856:  89%|████████▉ | 223/250 [07:35<00:54,  2.04s/it]\n",
      "Loss: 0.01828:  89%|████████▉ | 223/250 [07:35<00:54,  2.04s/it]\n",
      "Loss: 0.01849:  89%|████████▉ | 223/250 [07:35<00:54,  2.04s/it]\n",
      "Loss: 0.01880:  89%|████████▉ | 223/250 [07:35<00:54,  2.04s/it]\n",
      "Loss: 0.01838:  89%|████████▉ | 223/250 [07:35<00:54,  2.04s/it]\n",
      "Loss: 0.01640:  89%|████████▉ | 223/250 [07:35<00:54,  2.04s/it]\n",
      "Loss: 0.01729:  89%|████████▉ | 223/250 [07:36<00:54,  2.04s/it]\n",
      "Loss: 0.01580:  89%|████████▉ | 223/250 [07:36<00:54,  2.04s/it]\n",
      "Loss: 0.01599:  89%|████████▉ | 223/250 [07:36<00:54,  2.04s/it]\n",
      "Loss: 0.01719:  89%|████████▉ | 223/250 [07:36<00:54,  2.04s/it]\n",
      "Loss: 0.01526:  89%|████████▉ | 223/250 [07:36<00:54,  2.04s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:01<00:00, 64.02it/s]\u001b[A\n",
      "Loss: 0.01526:  90%|████████▉ | 224/250 [07:36<00:52,  2.03s/it]\n",
      "Loss: 0.01733:  90%|████████▉ | 224/250 [07:36<00:52,  2.03s/it]\n",
      "Loss: 0.01608:  90%|████████▉ | 224/250 [07:36<00:52,  2.03s/it]\n",
      "Loss: 0.01981:  90%|████████▉ | 224/250 [07:36<00:52,  2.03s/it]\n",
      "Loss: 0.01913:  90%|████████▉ | 224/250 [07:36<00:52,  2.03s/it]\n",
      "Loss: 0.01753:  90%|████████▉ | 224/250 [07:37<00:52,  2.03s/it]\n",
      "Loss: 0.01615:  90%|████████▉ | 224/250 [07:37<00:52,  2.03s/it]\n",
      "Loss: 0.01756:  90%|████████▉ | 224/250 [07:37<00:52,  2.03s/it]\n",
      "Loss: 0.01973:  90%|████████▉ | 224/250 [07:37<00:52,  2.03s/it]\n",
      "Loss: 0.01654:  90%|████████▉ | 224/250 [07:37<00:52,  2.03s/it]\n",
      "Loss: 0.01574:  90%|████████▉ | 224/250 [07:37<00:52,  2.03s/it]\n",
      "Loss: 0.01980:  90%|████████▉ | 224/250 [07:37<00:52,  2.03s/it]\n",
      "Loss: 0.01924:  90%|████████▉ | 224/250 [07:37<00:52,  2.03s/it]\n",
      "Loss: 0.01846:  90%|████████▉ | 224/250 [07:37<00:52,  2.03s/it]\n",
      "Loss: 0.01650:  90%|████████▉ | 224/250 [07:38<00:52,  2.03s/it]\n",
      "Loss: 0.01716:  90%|████████▉ | 224/250 [07:38<00:52,  2.03s/it]\n",
      "Loss: 0.01589:  90%|████████▉ | 224/250 [07:38<00:52,  2.03s/it]\n",
      "Loss: 0.01588:  90%|████████▉ | 224/250 [07:38<00:52,  2.03s/it]\n",
      "Loss: 0.01720:  90%|████████▉ | 224/250 [07:38<00:52,  2.03s/it]\n",
      "Loss: 0.01499:  90%|████████▉ | 224/250 [07:38<00:52,  2.03s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.88it/s]\u001b[A\n",
      "Loss: 0.01499:  90%|█████████ | 225/250 [07:38<00:50,  2.04s/it]\n",
      "Loss: 0.01690:  90%|█████████ | 225/250 [07:38<00:50,  2.04s/it]\n",
      "Loss: 0.01615:  90%|█████████ | 225/250 [07:38<00:50,  2.04s/it]\n",
      "Loss: 0.01630:  90%|█████████ | 225/250 [07:38<00:50,  2.04s/it]\n",
      "Loss: 0.01884:  90%|█████████ | 225/250 [07:39<00:50,  2.04s/it]\n",
      "Loss: 0.01493:  90%|█████████ | 225/250 [07:39<00:50,  2.04s/it]\n",
      "Loss: 0.01790:  90%|█████████ | 225/250 [07:39<00:50,  2.04s/it]\n",
      "Loss: 0.01962:  90%|█████████ | 225/250 [07:39<00:50,  2.04s/it]\n",
      "Loss: 0.01797:  90%|█████████ | 225/250 [07:39<00:50,  2.04s/it]\n",
      "Loss: 0.01850:  90%|█████████ | 225/250 [07:39<00:50,  2.04s/it]\n",
      "Loss: 0.01746:  90%|█████████ | 225/250 [07:39<00:50,  2.04s/it]\n",
      "Loss: 0.01608:  90%|█████████ | 225/250 [07:39<00:50,  2.04s/it]\n",
      "Loss: 0.01885:  90%|█████████ | 225/250 [07:39<00:50,  2.04s/it]\n",
      "Loss: 0.01647:  90%|█████████ | 225/250 [07:40<00:50,  2.04s/it]\n",
      "Loss: 0.01842:  90%|█████████ | 225/250 [07:40<00:50,  2.04s/it]\n",
      "Loss: 0.01566:  90%|█████████ | 225/250 [07:40<00:50,  2.04s/it]\n",
      "Loss: 0.01726:  90%|█████████ | 225/250 [07:40<00:50,  2.04s/it]\n",
      "Loss: 0.01478:  90%|█████████ | 225/250 [07:40<00:50,  2.04s/it]\n",
      "Loss: 0.01766:  90%|█████████ | 225/250 [07:40<00:50,  2.04s/it]\n",
      "Loss: 0.01507:  90%|█████████ | 225/250 [07:40<00:50,  2.04s/it]\n",
      "Loss: 0.01507:  90%|█████████ | 226/250 [07:40<00:48,  2.04s/it]\n",
      "Loss: 0.01694:  90%|█████████ | 226/250 [07:40<00:48,  2.04s/it]\n",
      "Loss: 0.01701:  90%|█████████ | 226/250 [07:40<00:48,  2.04s/it]\n",
      "Loss: 0.01927:  90%|█████████ | 226/250 [07:40<00:48,  2.04s/it]\n",
      "Loss: 0.01839:  90%|█████████ | 226/250 [07:41<00:48,  2.04s/it]\n",
      "Loss: 0.01505:  90%|█████████ | 226/250 [07:41<00:48,  2.04s/it]\n",
      "Loss: 0.01800:  90%|█████████ | 226/250 [07:41<00:48,  2.04s/it]\n",
      "Loss: 0.01591:  90%|█████████ | 226/250 [07:41<00:48,  2.04s/it]\n",
      "Loss: 0.01745:  90%|█████████ | 226/250 [07:41<00:48,  2.04s/it]\n",
      "Loss: 0.01844:  90%|█████████ | 226/250 [07:41<00:48,  2.04s/it]\n",
      "Loss: 0.01780:  90%|█████████ | 226/250 [07:41<00:48,  2.04s/it]\n",
      "Loss: 0.01656:  90%|█████████ | 226/250 [07:41<00:48,  2.04s/it]\n",
      "Loss: 0.01680:  90%|█████████ | 226/250 [07:41<00:48,  2.04s/it]\n",
      "Loss: 0.01755:  90%|█████████ | 226/250 [07:42<00:48,  2.04s/it]\n",
      "Loss: 0.01615:  90%|█████████ | 226/250 [07:42<00:48,  2.04s/it]\n",
      "Loss: 0.01769:  90%|█████████ | 226/250 [07:42<00:48,  2.04s/it]\n",
      "Loss: 0.01835:  90%|█████████ | 226/250 [07:42<00:48,  2.04s/it]\n",
      "Loss: 0.01897:  90%|█████████ | 226/250 [07:42<00:48,  2.04s/it]\n",
      "Loss: 0.01582:  90%|█████████ | 226/250 [07:42<00:48,  2.04s/it]\n",
      "Loss: 0.01522:  90%|█████████ | 226/250 [07:42<00:48,  2.04s/it]\n",
      "Loss: 0.01522:  91%|█████████ | 227/250 [07:42<00:46,  2.04s/it]\n",
      "Loss: 0.01728:  91%|█████████ | 227/250 [07:42<00:46,  2.04s/it]\n",
      "Loss: 0.01641:  91%|█████████ | 227/250 [07:42<00:46,  2.04s/it]\n",
      "Loss: 0.01956:  91%|█████████ | 227/250 [07:43<00:46,  2.04s/it]\n",
      "Loss: 0.01585:  91%|█████████ | 227/250 [07:43<00:46,  2.04s/it]\n",
      "Loss: 0.01716:  91%|█████████ | 227/250 [07:43<00:46,  2.04s/it]\n",
      "Loss: 0.01799:  91%|█████████ | 227/250 [07:43<00:46,  2.04s/it]\n",
      "Loss: 0.01587:  91%|█████████ | 227/250 [07:43<00:46,  2.04s/it]\n",
      "Loss: 0.01709:  91%|█████████ | 227/250 [07:43<00:46,  2.04s/it]\n",
      "Loss: 0.01777:  91%|█████████ | 227/250 [07:43<00:46,  2.04s/it]\n",
      "Loss: 0.01698:  91%|█████████ | 227/250 [07:43<00:46,  2.04s/it]\n",
      "Loss: 0.01967:  91%|█████████ | 227/250 [07:43<00:46,  2.04s/it]\n",
      "Loss: 0.01795:  91%|█████████ | 227/250 [07:44<00:46,  2.04s/it]\n",
      "Loss: 0.01751:  91%|█████████ | 227/250 [07:44<00:46,  2.04s/it]\n",
      "Loss: 0.01921:  91%|█████████ | 227/250 [07:44<00:46,  2.04s/it]\n",
      "Loss: 0.01721:  91%|█████████ | 227/250 [07:44<00:46,  2.04s/it]\n",
      "Loss: 0.01595:  91%|█████████ | 227/250 [07:44<00:46,  2.04s/it]\n",
      "Loss: 0.01815:  91%|█████████ | 227/250 [07:44<00:46,  2.04s/it]\n",
      "Loss: 0.01886:  91%|█████████ | 227/250 [07:44<00:46,  2.04s/it]\n",
      "Loss: 0.01504:  91%|█████████ | 227/250 [07:44<00:46,  2.04s/it]\n",
      "Loss: 0.01504:  91%|█████████ | 228/250 [07:44<00:45,  2.05s/it]\n",
      "Loss: 0.01731:  91%|█████████ | 228/250 [07:44<00:45,  2.05s/it]\n",
      "Loss: 0.01606:  91%|█████████ | 228/250 [07:44<00:45,  2.05s/it]\n",
      "Loss: 0.01618:  91%|█████████ | 228/250 [07:45<00:45,  2.05s/it]\n",
      "Loss: 0.01862:  91%|█████████ | 228/250 [07:45<00:45,  2.05s/it]\n",
      "Loss: 0.01525:  91%|█████████ | 228/250 [07:45<00:45,  2.05s/it]\n",
      "Loss: 0.01784:  91%|█████████ | 228/250 [07:45<00:45,  2.05s/it]\n",
      "Loss: 0.01573:  91%|█████████ | 228/250 [07:45<00:45,  2.05s/it]\n",
      "Loss: 0.01716:  91%|█████████ | 228/250 [07:45<00:45,  2.05s/it]\n",
      "Loss: 0.01800:  91%|█████████ | 228/250 [07:45<00:45,  2.05s/it]\n",
      "Loss: 0.01755:  91%|█████████ | 228/250 [07:45<00:45,  2.05s/it]\n",
      "Loss: 0.01633:  91%|█████████ | 228/250 [07:45<00:45,  2.05s/it]\n",
      "Loss: 0.01936:  91%|█████████ | 228/250 [07:46<00:45,  2.05s/it]\n",
      "Loss: 0.01648:  91%|█████████ | 228/250 [07:46<00:45,  2.05s/it]\n",
      "Loss: 0.01852:  91%|█████████ | 228/250 [07:46<00:45,  2.05s/it]\n",
      "Loss: 0.01535:  91%|█████████ | 228/250 [07:46<00:45,  2.05s/it]\n",
      "Loss: 0.01711:  91%|█████████ | 228/250 [07:46<00:45,  2.05s/it]\n",
      "Loss: 0.01478:  91%|█████████ | 228/250 [07:46<00:45,  2.05s/it]\n",
      "Loss: 0.01767:  91%|█████████ | 228/250 [07:46<00:45,  2.05s/it]\n",
      "Loss: 0.01513:  91%|█████████ | 228/250 [07:46<00:45,  2.05s/it]\n",
      "Loss: 0.01513:  92%|█████████▏| 229/250 [07:46<00:42,  2.04s/it]\n",
      "Loss: 0.01746:  92%|█████████▏| 229/250 [07:46<00:42,  2.04s/it]\n",
      "Loss: 0.01724:  92%|█████████▏| 229/250 [07:47<00:42,  2.04s/it]\n",
      "Loss: 0.01903:  92%|█████████▏| 229/250 [07:47<00:42,  2.04s/it]\n",
      "Loss: 0.01713:  92%|█████████▏| 229/250 [07:47<00:42,  2.04s/it]\n",
      "Loss: 0.01669:  92%|█████████▏| 229/250 [07:47<00:42,  2.04s/it]\n",
      "Loss: 0.01705:  92%|█████████▏| 229/250 [07:47<00:42,  2.04s/it]\n",
      "Loss: 0.01758:  92%|█████████▏| 229/250 [07:47<00:42,  2.04s/it]\n",
      "Loss: 0.01747:  92%|█████████▏| 229/250 [07:47<00:42,  2.04s/it]\n",
      "Loss: 0.01571:  92%|█████████▏| 229/250 [07:47<00:42,  2.04s/it]\n",
      "Loss: 0.01867:  92%|█████████▏| 229/250 [07:47<00:42,  2.04s/it]\n",
      "Loss: 0.01952:  92%|█████████▏| 229/250 [07:48<00:42,  2.04s/it]\n",
      "Loss: 0.01718:  92%|█████████▏| 229/250 [07:48<00:42,  2.04s/it]\n",
      "Loss: 0.01750:  92%|█████████▏| 229/250 [07:48<00:42,  2.04s/it]\n",
      "Loss: 0.01870:  92%|█████████▏| 229/250 [07:48<00:42,  2.04s/it]\n",
      "Loss: 0.01754:  92%|█████████▏| 229/250 [07:48<00:42,  2.04s/it]\n",
      "Loss: 0.01589:  92%|█████████▏| 229/250 [07:48<00:42,  2.04s/it]\n",
      "Loss: 0.01832:  92%|█████████▏| 229/250 [07:48<00:42,  2.04s/it]\n",
      "Loss: 0.01884:  92%|█████████▏| 229/250 [07:48<00:42,  2.04s/it]\n",
      "Loss: 0.01529:  92%|█████████▏| 229/250 [07:48<00:42,  2.04s/it]\n",
      "Loss: 0.01529:  92%|█████████▏| 230/250 [07:48<00:40,  2.04s/it]\n",
      "Loss: 0.01699:  92%|█████████▏| 230/250 [07:48<00:40,  2.04s/it]\n",
      "Loss: 0.01715:  92%|█████████▏| 230/250 [07:49<00:40,  2.04s/it]\n",
      "Loss: 0.01897:  92%|█████████▏| 230/250 [07:49<00:40,  2.04s/it]\n",
      "Loss: 0.01700:  92%|█████████▏| 230/250 [07:49<00:40,  2.04s/it]\n",
      "Loss: 0.01687:  92%|█████████▏| 230/250 [07:49<00:40,  2.04s/it]\n",
      "Loss: 0.01719:  92%|█████████▏| 230/250 [07:49<00:40,  2.04s/it]\n",
      "Loss: 0.01734:  92%|█████████▏| 230/250 [07:49<00:40,  2.04s/it]\n",
      "Loss: 0.01744:  92%|█████████▏| 230/250 [07:49<00:40,  2.04s/it]\n",
      "Loss: 0.01569:  92%|█████████▏| 230/250 [07:49<00:40,  2.04s/it]\n",
      "Loss: 0.01800:  92%|█████████▏| 230/250 [07:49<00:40,  2.04s/it]\n",
      "Loss: 0.01932:  92%|█████████▏| 230/250 [07:50<00:40,  2.04s/it]\n",
      "Loss: 0.01628:  92%|█████████▏| 230/250 [07:50<00:40,  2.04s/it]\n",
      "Loss: 0.01723:  92%|█████████▏| 230/250 [07:50<00:40,  2.04s/it]\n",
      "Loss: 0.01710:  92%|█████████▏| 230/250 [07:50<00:40,  2.04s/it]\n",
      "Loss: 0.01656:  92%|█████████▏| 230/250 [07:50<00:40,  2.04s/it]\n",
      "Loss: 0.01837:  92%|█████████▏| 230/250 [07:50<00:40,  2.04s/it]\n",
      "Loss: 0.01708:  92%|█████████▏| 230/250 [07:50<00:40,  2.04s/it]\n",
      "Loss: 0.01505:  92%|█████████▏| 230/250 [07:50<00:40,  2.04s/it]\n",
      "Loss: 0.01505:  92%|█████████▏| 231/250 [07:50<00:38,  2.04s/it]\n",
      "Loss: 0.01696:  92%|█████████▏| 231/250 [07:50<00:38,  2.04s/it]\n",
      "Loss: 0.01713:  92%|█████████▏| 231/250 [07:51<00:38,  2.04s/it]\n",
      "Loss: 0.01862:  92%|█████████▏| 231/250 [07:51<00:38,  2.04s/it]\n",
      "Loss: 0.01586:  92%|█████████▏| 231/250 [07:51<00:38,  2.04s/it]\n",
      "Loss: 0.01699:  92%|█████████▏| 231/250 [07:51<00:38,  2.04s/it]\n",
      "Loss: 0.01514:  92%|█████████▏| 231/250 [07:51<00:38,  2.04s/it]\n",
      "Loss: 0.01654:  92%|█████████▏| 231/250 [07:51<00:38,  2.04s/it]\n",
      "Loss: 0.01836:  92%|█████████▏| 231/250 [07:51<00:38,  2.04s/it]\n",
      "Loss: 0.01730:  92%|█████████▏| 231/250 [07:51<00:38,  2.04s/it]\n",
      "Loss: 0.01712:  92%|█████████▏| 231/250 [07:51<00:38,  2.04s/it]\n",
      "Loss: 0.02039:  92%|█████████▏| 231/250 [07:52<00:38,  2.04s/it]\n",
      "Loss: 0.01705:  92%|█████████▏| 231/250 [07:52<00:38,  2.04s/it]\n",
      "Loss: 0.01774:  92%|█████████▏| 231/250 [07:52<00:38,  2.04s/it]\n",
      "Loss: 0.01838:  92%|█████████▏| 231/250 [07:52<00:38,  2.04s/it]\n",
      "Loss: 0.01714:  92%|█████████▏| 231/250 [07:52<00:38,  2.04s/it]\n",
      "Loss: 0.01600:  92%|█████████▏| 231/250 [07:52<00:38,  2.04s/it]\n",
      "Loss: 0.01858:  92%|█████████▏| 231/250 [07:52<00:38,  2.04s/it]\n",
      "Loss: 0.01872:  92%|█████████▏| 231/250 [07:52<00:38,  2.04s/it]\n",
      "Loss: 0.01525:  92%|█████████▏| 231/250 [07:52<00:38,  2.04s/it]\n",
      "Loss: 0.01525:  93%|█████████▎| 232/250 [07:52<00:36,  2.04s/it]\n",
      "Loss: 0.01693:  93%|█████████▎| 232/250 [07:53<00:36,  2.04s/it]\n",
      "Loss: 0.01612:  93%|█████████▎| 232/250 [07:53<00:36,  2.04s/it]\n",
      "Loss: 0.01919:  93%|█████████▎| 232/250 [07:53<00:36,  2.04s/it]\n",
      "Loss: 0.01871:  93%|█████████▎| 232/250 [07:53<00:36,  2.04s/it]\n",
      "Loss: 0.01484:  93%|█████████▎| 232/250 [07:53<00:36,  2.04s/it]\n",
      "Loss: 0.01759:  93%|█████████▎| 232/250 [07:53<00:36,  2.04s/it]\n",
      "Loss: 0.01948:  93%|█████████▎| 232/250 [07:53<00:36,  2.04s/it]\n",
      "Loss: 0.01867:  93%|█████████▎| 232/250 [07:53<00:36,  2.04s/it]\n",
      "Loss: 0.01845:  93%|█████████▎| 232/250 [07:53<00:36,  2.04s/it]\n",
      "Loss: 0.01749:  93%|█████████▎| 232/250 [07:54<00:36,  2.04s/it]\n",
      "Loss: 0.01623:  93%|█████████▎| 232/250 [07:54<00:36,  2.04s/it]\n",
      "Loss: 0.01894:  93%|█████████▎| 232/250 [07:54<00:36,  2.04s/it]\n",
      "Loss: 0.01648:  93%|█████████▎| 232/250 [07:54<00:36,  2.04s/it]\n",
      "Loss: 0.01822:  93%|█████████▎| 232/250 [07:54<00:36,  2.04s/it]\n",
      "Loss: 0.01532:  93%|█████████▎| 232/250 [07:54<00:36,  2.04s/it]\n",
      "Loss: 0.01668:  93%|█████████▎| 232/250 [07:54<00:36,  2.04s/it]\n",
      "Loss: 0.01474:  93%|█████████▎| 232/250 [07:54<00:36,  2.04s/it]\n",
      "Loss: 0.01773:  93%|█████████▎| 232/250 [07:54<00:36,  2.04s/it]\n",
      "Loss: 0.01509:  93%|█████████▎| 232/250 [07:54<00:36,  2.04s/it]\n",
      "Loss: 0.01509:  93%|█████████▎| 233/250 [07:54<00:34,  2.05s/it]\n",
      "Loss: 0.01708:  93%|█████████▎| 233/250 [07:55<00:34,  2.05s/it]\n",
      "Loss: 0.01756:  93%|█████████▎| 233/250 [07:55<00:34,  2.05s/it]\n",
      "Loss: 0.01910:  93%|█████████▎| 233/250 [07:55<00:34,  2.05s/it]\n",
      "Loss: 0.01827:  93%|█████████▎| 233/250 [07:55<00:34,  2.05s/it]\n",
      "Loss: 0.01746:  93%|█████████▎| 233/250 [07:55<00:34,  2.05s/it]\n",
      "Loss: 0.01765:  93%|█████████▎| 233/250 [07:55<00:34,  2.05s/it]\n",
      "Loss: 0.01897:  93%|█████████▎| 233/250 [07:55<00:34,  2.05s/it]\n",
      "Loss: 0.01738:  93%|█████████▎| 233/250 [07:55<00:34,  2.05s/it]\n",
      "Loss: 0.01779:  93%|█████████▎| 233/250 [07:55<00:34,  2.05s/it]\n",
      "Loss: 0.01699:  93%|█████████▎| 233/250 [07:56<00:34,  2.05s/it]\n",
      "Loss: 0.01658:  93%|█████████▎| 233/250 [07:56<00:34,  2.05s/it]\n",
      "Loss: 0.01574:  93%|█████████▎| 233/250 [07:56<00:34,  2.05s/it]\n",
      "Loss: 0.01991:  93%|█████████▎| 233/250 [07:56<00:34,  2.05s/it]\n",
      "Loss: 0.01772:  93%|█████████▎| 233/250 [07:56<00:34,  2.05s/it]\n",
      "Loss: 0.01670:  93%|█████████▎| 233/250 [07:56<00:34,  2.05s/it]\n",
      "Loss: 0.02011:  93%|█████████▎| 233/250 [07:56<00:34,  2.05s/it]\n",
      "Loss: 0.01524:  93%|█████████▎| 233/250 [07:56<00:34,  2.05s/it]\n",
      "Loss: 0.01704:  93%|█████████▎| 233/250 [07:56<00:34,  2.05s/it]\n",
      "Loss: 0.01493:  93%|█████████▎| 233/250 [07:57<00:34,  2.05s/it]\n",
      "Loss: 0.01493:  94%|█████████▎| 234/250 [07:57<00:32,  2.06s/it]\n",
      "Loss: 0.01758:  94%|█████████▎| 234/250 [07:57<00:32,  2.06s/it]\n",
      "Loss: 0.01685:  94%|█████████▎| 234/250 [07:57<00:32,  2.06s/it]\n",
      "Loss: 0.01878:  94%|█████████▎| 234/250 [07:57<00:32,  2.06s/it]\n",
      "Loss: 0.01747:  94%|█████████▎| 234/250 [07:57<00:32,  2.06s/it]\n",
      "Loss: 0.01660:  94%|█████████▎| 234/250 [07:57<00:32,  2.06s/it]\n",
      "Loss: 0.01597:  94%|█████████▎| 234/250 [07:57<00:32,  2.06s/it]\n",
      "Loss: 0.01677:  94%|█████████▎| 234/250 [07:57<00:32,  2.06s/it]\n",
      "Loss: 0.01670:  94%|█████████▎| 234/250 [07:57<00:32,  2.06s/it]\n",
      "Loss: 0.01805:  94%|█████████▎| 234/250 [07:58<00:32,  2.06s/it]\n",
      "Loss: 0.01700:  94%|█████████▎| 234/250 [07:58<00:32,  2.06s/it]\n",
      "Loss: 0.01936:  94%|█████████▎| 234/250 [07:58<00:32,  2.06s/it]\n",
      "Loss: 0.01727:  94%|█████████▎| 234/250 [07:58<00:32,  2.06s/it]\n",
      "Loss: 0.01781:  94%|█████████▎| 234/250 [07:58<00:32,  2.06s/it]\n",
      "Loss: 0.01847:  94%|█████████▎| 234/250 [07:58<00:32,  2.06s/it]\n",
      "Loss: 0.01736:  94%|█████████▎| 234/250 [07:58<00:32,  2.06s/it]\n",
      "Loss: 0.01592:  94%|█████████▎| 234/250 [07:58<00:32,  2.06s/it]\n",
      "Loss: 0.01861:  94%|█████████▎| 234/250 [07:58<00:32,  2.06s/it]\n",
      "Loss: 0.01869:  94%|█████████▎| 234/250 [07:59<00:32,  2.06s/it]\n",
      "Loss: 0.01557:  94%|█████████▎| 234/250 [07:59<00:32,  2.06s/it]\n",
      "Loss: 0.01557:  94%|█████████▍| 235/250 [07:59<00:30,  2.05s/it]\n",
      "Loss: 0.01698:  94%|█████████▍| 235/250 [07:59<00:30,  2.05s/it]\n",
      "Loss: 0.01721:  94%|█████████▍| 235/250 [07:59<00:30,  2.05s/it]\n",
      "Loss: 0.01864:  94%|█████████▍| 235/250 [07:59<00:30,  2.05s/it]\n",
      "Loss: 0.01708:  94%|█████████▍| 235/250 [07:59<00:30,  2.05s/it]\n",
      "Loss: 0.01653:  94%|█████████▍| 235/250 [07:59<00:30,  2.05s/it]\n",
      "Loss: 0.01509:  94%|█████████▍| 235/250 [07:59<00:30,  2.05s/it]\n",
      "Loss: 0.01686:  94%|█████████▍| 235/250 [07:59<00:30,  2.05s/it]\n",
      "Loss: 0.01690:  94%|█████████▍| 235/250 [07:59<00:30,  2.05s/it]\n",
      "Loss: 0.01783:  94%|█████████▍| 235/250 [08:00<00:30,  2.05s/it]\n",
      "Loss: 0.01746:  94%|█████████▍| 235/250 [08:00<00:30,  2.05s/it]\n",
      "Loss: 0.01605:  94%|█████████▍| 235/250 [08:00<00:30,  2.05s/it]\n",
      "Loss: 0.01572:  94%|█████████▍| 235/250 [08:00<00:30,  2.05s/it]\n",
      "Loss: 0.02045:  94%|█████████▍| 235/250 [08:00<00:30,  2.05s/it]\n",
      "Loss: 0.01753:  94%|█████████▍| 235/250 [08:00<00:30,  2.05s/it]\n",
      "Loss: 0.01692:  94%|█████████▍| 235/250 [08:00<00:30,  2.05s/it]\n",
      "Loss: 0.02035:  94%|█████████▍| 235/250 [08:00<00:30,  2.05s/it]\n",
      "Loss: 0.01530:  94%|█████████▍| 235/250 [08:00<00:30,  2.05s/it]\n",
      "Loss: 0.01758:  94%|█████████▍| 235/250 [08:01<00:30,  2.05s/it]\n",
      "Loss: 0.01488:  94%|█████████▍| 235/250 [08:01<00:30,  2.05s/it]\n",
      "Loss: 0.01488:  94%|█████████▍| 236/250 [08:01<00:28,  2.05s/it]\n",
      "Loss: 0.01704:  94%|█████████▍| 236/250 [08:01<00:28,  2.05s/it]\n",
      "Loss: 0.01703:  94%|█████████▍| 236/250 [08:01<00:28,  2.05s/it]\n",
      "Loss: 0.01865:  94%|█████████▍| 236/250 [08:01<00:28,  2.05s/it]\n",
      "Loss: 0.01752:  94%|█████████▍| 236/250 [08:01<00:28,  2.05s/it]\n",
      "Loss: 0.01690:  94%|█████████▍| 236/250 [08:01<00:28,  2.05s/it]\n",
      "Loss: 0.01697:  94%|█████████▍| 236/250 [08:01<00:28,  2.05s/it]\n",
      "Loss: 0.01657:  94%|█████████▍| 236/250 [08:01<00:28,  2.05s/it]\n",
      "Loss: 0.01691:  94%|█████████▍| 236/250 [08:02<00:28,  2.05s/it]\n",
      "Loss: 0.01775:  94%|█████████▍| 236/250 [08:02<00:28,  2.05s/it]\n",
      "Loss: 0.01767:  94%|█████████▍| 236/250 [08:02<00:28,  2.05s/it]\n",
      "Loss: 0.01587:  94%|█████████▍| 236/250 [08:02<00:28,  2.05s/it]\n",
      "Loss: 0.01882:  94%|█████████▍| 236/250 [08:02<00:28,  2.05s/it]\n",
      "Loss: 0.01661:  94%|█████████▍| 236/250 [08:02<00:28,  2.05s/it]\n",
      "Loss: 0.01818:  94%|█████████▍| 236/250 [08:02<00:28,  2.05s/it]\n",
      "Loss: 0.01586:  94%|█████████▍| 236/250 [08:02<00:28,  2.05s/it]\n",
      "Loss: 0.01673:  94%|█████████▍| 236/250 [08:02<00:28,  2.05s/it]\n",
      "Loss: 0.01461:  94%|█████████▍| 236/250 [08:03<00:28,  2.05s/it]\n",
      "Loss: 0.01726:  94%|█████████▍| 236/250 [08:03<00:28,  2.05s/it]\n",
      "Loss: 0.01506:  94%|█████████▍| 236/250 [08:03<00:28,  2.05s/it]\n",
      "Loss: 0.01506:  95%|█████████▍| 237/250 [08:03<00:26,  2.05s/it]\n",
      "Loss: 0.01772:  95%|█████████▍| 237/250 [08:03<00:26,  2.05s/it]\n",
      "Loss: 0.01713:  95%|█████████▍| 237/250 [08:03<00:26,  2.05s/it]\n",
      "Loss: 0.01881:  95%|█████████▍| 237/250 [08:03<00:26,  2.05s/it]\n",
      "Loss: 0.01704:  95%|█████████▍| 237/250 [08:03<00:26,  2.05s/it]\n",
      "Loss: 0.01646:  95%|█████████▍| 237/250 [08:03<00:26,  2.05s/it]\n",
      "Loss: 0.01678:  95%|█████████▍| 237/250 [08:03<00:26,  2.05s/it]\n",
      "Loss: 0.01762:  95%|█████████▍| 237/250 [08:03<00:26,  2.05s/it]\n",
      "Loss: 0.01748:  95%|█████████▍| 237/250 [08:04<00:26,  2.05s/it]\n",
      "Loss: 0.01566:  95%|█████████▍| 237/250 [08:04<00:26,  2.05s/it]\n",
      "Loss: 0.01797:  95%|█████████▍| 237/250 [08:04<00:26,  2.05s/it]\n",
      "Loss: 0.01927:  95%|█████████▍| 237/250 [08:04<00:26,  2.05s/it]\n",
      "Loss: 0.01685:  95%|█████████▍| 237/250 [08:04<00:26,  2.05s/it]\n",
      "Loss: 0.01694:  95%|█████████▍| 237/250 [08:04<00:26,  2.05s/it]\n",
      "Loss: 0.01692:  95%|█████████▍| 237/250 [08:04<00:26,  2.05s/it]\n",
      "Loss: 0.01687:  95%|█████████▍| 237/250 [08:04<00:26,  2.05s/it]\n",
      "Loss: 0.01773:  95%|█████████▍| 237/250 [08:05<00:26,  2.05s/it]\n",
      "Loss: 0.01758:  95%|█████████▍| 237/250 [08:05<00:26,  2.05s/it]\n",
      "Loss: 0.01482:  95%|█████████▍| 237/250 [08:05<00:26,  2.05s/it]\n",
      "Loss: 0.01482:  95%|█████████▌| 238/250 [08:05<00:24,  2.05s/it]\n",
      "Loss: 0.01703:  95%|█████████▌| 238/250 [08:05<00:24,  2.05s/it]\n",
      "Loss: 0.01647:  95%|█████████▌| 238/250 [08:05<00:24,  2.05s/it]\n",
      "Loss: 0.01909:  95%|█████████▌| 238/250 [08:05<00:24,  2.05s/it]\n",
      "Loss: 0.01605:  95%|█████████▌| 238/250 [08:05<00:24,  2.05s/it]\n",
      "Loss: 0.01694:  95%|█████████▌| 238/250 [08:05<00:24,  2.05s/it]\n",
      "Loss: 0.01498:  95%|█████████▌| 238/250 [08:05<00:24,  2.05s/it]\n",
      "Loss: 0.01577:  95%|█████████▌| 238/250 [08:06<00:24,  2.05s/it]\n",
      "Loss: 0.01665:  95%|█████████▌| 238/250 [08:06<00:24,  2.05s/it]\n",
      "Loss: 0.01807:  95%|█████████▌| 238/250 [08:06<00:24,  2.05s/it]\n",
      "Loss: 0.01719:  95%|█████████▌| 238/250 [08:06<00:24,  2.05s/it]\n",
      "Loss: 0.01654:  95%|█████████▌| 238/250 [08:06<00:24,  2.05s/it]\n",
      "Loss: 0.01925:  95%|█████████▌| 238/250 [08:06<00:24,  2.05s/it]\n",
      "Loss: 0.01646:  95%|█████████▌| 238/250 [08:06<00:24,  2.05s/it]\n",
      "Loss: 0.01809:  95%|█████████▌| 238/250 [08:06<00:24,  2.05s/it]\n",
      "Loss: 0.01537:  95%|█████████▌| 238/250 [08:06<00:24,  2.05s/it]\n",
      "Loss: 0.01710:  95%|█████████▌| 238/250 [08:07<00:24,  2.05s/it]\n",
      "Loss: 0.01455:  95%|█████████▌| 238/250 [08:07<00:24,  2.05s/it]\n",
      "Loss: 0.01730:  95%|█████████▌| 238/250 [08:07<00:24,  2.05s/it]\n",
      "Loss: 0.01522:  95%|█████████▌| 238/250 [08:07<00:24,  2.05s/it]\n",
      "Loss: 0.01522:  96%|█████████▌| 239/250 [08:07<00:22,  2.06s/it]\n",
      "Loss: 0.01705:  96%|█████████▌| 239/250 [08:07<00:22,  2.06s/it]\n",
      "Loss: 0.01622:  96%|█████████▌| 239/250 [08:07<00:22,  2.06s/it]\n",
      "Loss: 0.01906:  96%|█████████▌| 239/250 [08:07<00:22,  2.06s/it]\n",
      "Loss: 0.01628:  96%|█████████▌| 239/250 [08:07<00:22,  2.06s/it]\n",
      "Loss: 0.01491:  96%|█████████▌| 239/250 [08:07<00:22,  2.06s/it]\n",
      "Loss: 0.01852:  96%|█████████▌| 239/250 [08:07<00:22,  2.06s/it]\n",
      "Loss: 0.01561:  96%|█████████▌| 239/250 [08:08<00:22,  2.06s/it]\n",
      "Loss: 0.01691:  96%|█████████▌| 239/250 [08:08<00:22,  2.06s/it]\n",
      "Loss: 0.01812:  96%|█████████▌| 239/250 [08:08<00:22,  2.06s/it]\n",
      "Loss: 0.01714:  96%|█████████▌| 239/250 [08:08<00:22,  2.06s/it]\n",
      "Loss: 0.01611:  96%|█████████▌| 239/250 [08:08<00:22,  2.06s/it]\n",
      "Loss: 0.01889:  96%|█████████▌| 239/250 [08:08<00:22,  2.06s/it]\n",
      "Loss: 0.01979:  96%|█████████▌| 239/250 [08:08<00:22,  2.06s/it]\n",
      "Loss: 0.01743:  96%|█████████▌| 239/250 [08:08<00:22,  2.06s/it]\n",
      "Loss: 0.01656:  96%|█████████▌| 239/250 [08:08<00:22,  2.06s/it]\n",
      "Loss: 0.01684:  96%|█████████▌| 239/250 [08:09<00:22,  2.06s/it]\n",
      "Loss: 0.01863:  96%|█████████▌| 239/250 [08:09<00:22,  2.06s/it]\n",
      "Loss: 0.01719:  96%|█████████▌| 239/250 [08:09<00:22,  2.06s/it]\n",
      "Loss: 0.01494:  96%|█████████▌| 239/250 [08:09<00:22,  2.06s/it]\n",
      "Loss: 0.01494:  96%|█████████▌| 240/250 [08:09<00:20,  2.06s/it]\n",
      "Loss: 0.01708:  96%|█████████▌| 240/250 [08:09<00:20,  2.06s/it]\n",
      "Loss: 0.01730:  96%|█████████▌| 240/250 [08:09<00:20,  2.06s/it]\n",
      "Loss: 0.01887:  96%|█████████▌| 240/250 [08:09<00:20,  2.06s/it]\n",
      "Loss: 0.01712:  96%|█████████▌| 240/250 [08:09<00:20,  2.06s/it]\n",
      "Loss: 0.01709:  96%|█████████▌| 240/250 [08:09<00:20,  2.06s/it]\n",
      "Loss: 0.01589:  96%|█████████▌| 240/250 [08:10<00:20,  2.06s/it]\n",
      "Loss: 0.01628:  96%|█████████▌| 240/250 [08:10<00:20,  2.06s/it]\n",
      "Loss: 0.01966:  96%|█████████▌| 240/250 [08:10<00:20,  2.06s/it]\n",
      "Loss: 0.01736:  96%|█████████▌| 240/250 [08:10<00:20,  2.06s/it]\n",
      "Loss: 0.01825:  96%|█████████▌| 240/250 [08:10<00:20,  2.06s/it]\n",
      "Loss: 0.01836:  96%|█████████▌| 240/250 [08:10<00:20,  2.06s/it]\n",
      "Loss: 0.01811:  96%|█████████▌| 240/250 [08:10<00:20,  2.06s/it]\n",
      "Loss: 0.01646:  96%|█████████▌| 240/250 [08:10<00:20,  2.06s/it]\n",
      "Loss: 0.01656:  96%|█████████▌| 240/250 [08:10<00:20,  2.06s/it]\n",
      "Loss: 0.01606:  96%|█████████▌| 240/250 [08:11<00:20,  2.06s/it]\n",
      "Loss: 0.01593:  96%|█████████▌| 240/250 [08:11<00:20,  2.06s/it]\n",
      "Loss: 0.01691:  96%|█████████▌| 240/250 [08:11<00:20,  2.06s/it]\n",
      "Loss: 0.01500:  96%|█████████▌| 240/250 [08:11<00:20,  2.06s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.60it/s]\u001b[A\n",
      "Loss: 0.01500:  96%|█████████▋| 241/250 [08:11<00:18,  2.05s/it]\n",
      "Loss: 0.01714:  96%|█████████▋| 241/250 [08:11<00:18,  2.05s/it]\n",
      "Loss: 0.01599:  96%|█████████▋| 241/250 [08:11<00:18,  2.05s/it]\n",
      "Loss: 0.01903:  96%|█████████▋| 241/250 [08:11<00:18,  2.05s/it]\n",
      "Loss: 0.01594:  96%|█████████▋| 241/250 [08:11<00:18,  2.05s/it]\n",
      "Loss: 0.01723:  96%|█████████▋| 241/250 [08:11<00:18,  2.05s/it]\n",
      "Loss: 0.01753:  96%|█████████▋| 241/250 [08:12<00:18,  2.05s/it]\n",
      "Loss: 0.01557:  96%|█████████▋| 241/250 [08:12<00:18,  2.05s/it]\n",
      "Loss: 0.01753:  96%|█████████▋| 241/250 [08:12<00:18,  2.05s/it]\n",
      "Loss: 0.01774:  96%|█████████▋| 241/250 [08:12<00:18,  2.05s/it]\n",
      "Loss: 0.01753:  96%|█████████▋| 241/250 [08:12<00:18,  2.05s/it]\n",
      "Loss: 0.01643:  96%|█████████▋| 241/250 [08:12<00:18,  2.05s/it]\n",
      "Loss: 0.01688:  96%|█████████▋| 241/250 [08:12<00:18,  2.05s/it]\n",
      "Loss: 0.01770:  96%|█████████▋| 241/250 [08:12<00:18,  2.05s/it]\n",
      "Loss: 0.01552:  96%|█████████▋| 241/250 [08:12<00:18,  2.05s/it]\n",
      "Loss: 0.01728:  96%|█████████▋| 241/250 [08:13<00:18,  2.05s/it]\n",
      "Loss: 0.01793:  96%|█████████▋| 241/250 [08:13<00:18,  2.05s/it]\n",
      "Loss: 0.01865:  96%|█████████▋| 241/250 [08:13<00:18,  2.05s/it]\n",
      "Loss: 0.01564:  96%|█████████▋| 241/250 [08:13<00:18,  2.05s/it]\n",
      "Loss: 0.01473:  96%|█████████▋| 241/250 [08:13<00:18,  2.05s/it]\n",
      "Loss: 0.01473:  97%|█████████▋| 242/250 [08:13<00:16,  2.05s/it]\n",
      "Loss: 0.01664:  97%|█████████▋| 242/250 [08:13<00:16,  2.05s/it]\n",
      "Loss: 0.01573:  97%|█████████▋| 242/250 [08:13<00:16,  2.05s/it]\n",
      "Loss: 0.01596:  97%|█████████▋| 242/250 [08:13<00:16,  2.05s/it]\n",
      "Loss: 0.01675:  97%|█████████▋| 242/250 [08:13<00:16,  2.05s/it]\n",
      "Loss: 0.01743:  97%|█████████▋| 242/250 [08:13<00:16,  2.05s/it]\n",
      "Loss: 0.01606:  97%|█████████▋| 242/250 [08:14<00:16,  2.05s/it]\n",
      "Loss: 0.01753:  97%|█████████▋| 242/250 [08:14<00:16,  2.05s/it]\n",
      "Loss: 0.01967:  97%|█████████▋| 242/250 [08:14<00:16,  2.05s/it]\n",
      "Loss: 0.01678:  97%|█████████▋| 242/250 [08:14<00:16,  2.05s/it]\n",
      "Loss: 0.01848:  97%|█████████▋| 242/250 [08:14<00:16,  2.05s/it]\n",
      "Loss: 0.01814:  97%|█████████▋| 242/250 [08:14<00:16,  2.05s/it]\n",
      "Loss: 0.01925:  97%|█████████▋| 242/250 [08:14<00:16,  2.05s/it]\n",
      "Loss: 0.01821:  97%|█████████▋| 242/250 [08:14<00:16,  2.05s/it]\n",
      "Loss: 0.01612:  97%|█████████▋| 242/250 [08:14<00:16,  2.05s/it]\n",
      "Loss: 0.01687:  97%|█████████▋| 242/250 [08:15<00:16,  2.05s/it]\n",
      "Loss: 0.01595:  97%|█████████▋| 242/250 [08:15<00:16,  2.05s/it]\n",
      "Loss: 0.01559:  97%|█████████▋| 242/250 [08:15<00:16,  2.05s/it]\n",
      "Loss: 0.01695:  97%|█████████▋| 242/250 [08:15<00:16,  2.05s/it]\n",
      "Loss: 0.01526:  97%|█████████▋| 242/250 [08:15<00:16,  2.05s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 60.69it/s]\u001b[A\n",
      "Loss: 0.01526:  97%|█████████▋| 243/250 [08:15<00:14,  2.06s/it]\n",
      "Loss: 0.01663:  97%|█████████▋| 243/250 [08:15<00:14,  2.06s/it]\n",
      "Loss: 0.01593:  97%|█████████▋| 243/250 [08:15<00:14,  2.06s/it]\n",
      "Loss: 0.01890:  97%|█████████▋| 243/250 [08:15<00:14,  2.06s/it]\n",
      "Loss: 0.01623:  97%|█████████▋| 243/250 [08:15<00:14,  2.06s/it]\n",
      "Loss: 0.01531:  97%|█████████▋| 243/250 [08:16<00:14,  2.06s/it]\n",
      "Loss: 0.01820:  97%|█████████▋| 243/250 [08:16<00:14,  2.06s/it]\n",
      "Loss: 0.01924:  97%|█████████▋| 243/250 [08:16<00:14,  2.06s/it]\n",
      "Loss: 0.01750:  97%|█████████▋| 243/250 [08:16<00:14,  2.06s/it]\n",
      "Loss: 0.01907:  97%|█████████▋| 243/250 [08:16<00:14,  2.06s/it]\n",
      "Loss: 0.01726:  97%|█████████▋| 243/250 [08:16<00:14,  2.06s/it]\n",
      "Loss: 0.01663:  97%|█████████▋| 243/250 [08:16<00:14,  2.06s/it]\n",
      "Loss: 0.01913:  97%|█████████▋| 243/250 [08:16<00:14,  2.06s/it]\n",
      "Loss: 0.01665:  97%|█████████▋| 243/250 [08:16<00:14,  2.06s/it]\n",
      "Loss: 0.01830:  97%|█████████▋| 243/250 [08:17<00:14,  2.06s/it]\n",
      "Loss: 0.01533:  97%|█████████▋| 243/250 [08:17<00:14,  2.06s/it]\n",
      "Loss: 0.01679:  97%|█████████▋| 243/250 [08:17<00:14,  2.06s/it]\n",
      "Loss: 0.01456:  97%|█████████▋| 243/250 [08:17<00:14,  2.06s/it]\n",
      "Loss: 0.01717:  97%|█████████▋| 243/250 [08:17<00:14,  2.06s/it]\n",
      "Loss: 0.01519:  97%|█████████▋| 243/250 [08:17<00:14,  2.06s/it]\n",
      "Loss: 0.01519:  98%|█████████▊| 244/250 [08:17<00:12,  2.06s/it]\n",
      "Loss: 0.01703:  98%|█████████▊| 244/250 [08:17<00:12,  2.06s/it]\n",
      "Loss: 0.01702:  98%|█████████▊| 244/250 [08:17<00:12,  2.06s/it]\n",
      "Loss: 0.01835:  98%|█████████▊| 244/250 [08:17<00:12,  2.06s/it]\n",
      "Loss: 0.01680:  98%|█████████▊| 244/250 [08:18<00:12,  2.06s/it]\n",
      "Loss: 0.01652:  98%|█████████▊| 244/250 [08:18<00:12,  2.06s/it]\n",
      "Loss: 0.01693:  98%|█████████▊| 244/250 [08:18<00:12,  2.06s/it]\n",
      "Loss: 0.01739:  98%|█████████▊| 244/250 [08:18<00:12,  2.06s/it]\n",
      "Loss: 0.01851:  98%|█████████▊| 244/250 [08:18<00:12,  2.06s/it]\n",
      "Loss: 0.01763:  98%|█████████▊| 244/250 [08:18<00:12,  2.06s/it]\n",
      "Loss: 0.01699:  98%|█████████▊| 244/250 [08:18<00:12,  2.06s/it]\n",
      "Loss: 0.01630:  98%|█████████▊| 244/250 [08:18<00:12,  2.06s/it]\n",
      "Loss: 0.01691:  98%|█████████▊| 244/250 [08:18<00:12,  2.06s/it]\n",
      "Loss: 0.01806:  98%|█████████▊| 244/250 [08:19<00:12,  2.06s/it]\n",
      "Loss: 0.01515:  98%|█████████▊| 244/250 [08:19<00:12,  2.06s/it]\n",
      "Loss: 0.01737:  98%|█████████▊| 244/250 [08:19<00:12,  2.06s/it]\n",
      "Loss: 0.01816:  98%|█████████▊| 244/250 [08:19<00:12,  2.06s/it]\n",
      "Loss: 0.01889:  98%|█████████▊| 244/250 [08:19<00:12,  2.06s/it]\n",
      "Loss: 0.01615:  98%|█████████▊| 244/250 [08:19<00:12,  2.06s/it]\n",
      "Loss: 0.01524:  98%|█████████▊| 244/250 [08:19<00:12,  2.06s/it]\n",
      "Loss: 0.01524:  98%|█████████▊| 245/250 [08:19<00:10,  2.05s/it]\n",
      "Loss: 0.01709:  98%|█████████▊| 245/250 [08:19<00:10,  2.05s/it]\n",
      "Loss: 0.01699:  98%|█████████▊| 245/250 [08:19<00:10,  2.05s/it]\n",
      "Loss: 0.01845:  98%|█████████▊| 245/250 [08:19<00:10,  2.05s/it]\n",
      "Loss: 0.01693:  98%|█████████▊| 245/250 [08:20<00:10,  2.05s/it]\n",
      "Loss: 0.01685:  98%|█████████▊| 245/250 [08:20<00:10,  2.05s/it]\n",
      "Loss: 0.01764:  98%|█████████▊| 245/250 [08:20<00:10,  2.05s/it]\n",
      "Loss: 0.01598:  98%|█████████▊| 245/250 [08:20<00:10,  2.05s/it]\n",
      "Loss: 0.01760:  98%|█████████▊| 245/250 [08:20<00:10,  2.05s/it]\n",
      "Loss: 0.01789:  98%|█████████▊| 245/250 [08:20<00:10,  2.05s/it]\n",
      "Loss: 0.01805:  98%|█████████▊| 245/250 [08:20<00:10,  2.05s/it]\n",
      "Loss: 0.01650:  98%|█████████▊| 245/250 [08:20<00:10,  2.05s/it]\n",
      "Loss: 0.01685:  98%|█████████▊| 245/250 [08:20<00:10,  2.05s/it]\n",
      "Loss: 0.01762:  98%|█████████▊| 245/250 [08:21<00:10,  2.05s/it]\n",
      "Loss: 0.01506:  98%|█████████▊| 245/250 [08:21<00:10,  2.05s/it]\n",
      "Loss: 0.01766:  98%|█████████▊| 245/250 [08:21<00:10,  2.05s/it]\n",
      "Loss: 0.01817:  98%|█████████▊| 245/250 [08:21<00:10,  2.05s/it]\n",
      "Loss: 0.01886:  98%|█████████▊| 245/250 [08:21<00:10,  2.05s/it]\n",
      "Loss: 0.01565:  98%|█████████▊| 245/250 [08:21<00:10,  2.05s/it]\n",
      "Loss: 0.01481:  98%|█████████▊| 245/250 [08:21<00:10,  2.05s/it]\n",
      "Loss: 0.01481:  98%|█████████▊| 246/250 [08:21<00:08,  2.04s/it]\n",
      "Loss: 0.01693:  98%|█████████▊| 246/250 [08:21<00:08,  2.04s/it]\n",
      "Loss: 0.01676:  98%|█████████▊| 246/250 [08:21<00:08,  2.04s/it]\n",
      "Loss: 0.01854:  98%|█████████▊| 246/250 [08:21<00:08,  2.04s/it]\n",
      "Loss: 0.01626:  98%|█████████▊| 246/250 [08:22<00:08,  2.04s/it]\n",
      "Loss: 0.01696:  98%|█████████▊| 246/250 [08:22<00:08,  2.04s/it]\n",
      "Loss: 0.01765:  98%|█████████▊| 246/250 [08:22<00:08,  2.04s/it]\n",
      "Loss: 0.01904:  98%|█████████▊| 246/250 [08:22<00:08,  2.04s/it]\n",
      "Loss: 0.01840:  98%|█████████▊| 246/250 [08:22<00:08,  2.04s/it]\n",
      "Loss: 0.01875:  98%|█████████▊| 246/250 [08:22<00:08,  2.04s/it]\n",
      "Loss: 0.01763:  98%|█████████▊| 246/250 [08:22<00:08,  2.04s/it]\n",
      "Loss: 0.01638:  98%|█████████▊| 246/250 [08:22<00:08,  2.04s/it]\n",
      "Loss: 0.01893:  98%|█████████▊| 246/250 [08:22<00:08,  2.04s/it]\n",
      "Loss: 0.01638:  98%|█████████▊| 246/250 [08:23<00:08,  2.04s/it]\n",
      "Loss: 0.01867:  98%|█████████▊| 246/250 [08:23<00:08,  2.04s/it]\n",
      "Loss: 0.01576:  98%|█████████▊| 246/250 [08:23<00:08,  2.04s/it]\n",
      "Loss: 0.01696:  98%|█████████▊| 246/250 [08:23<00:08,  2.04s/it]\n",
      "Loss: 0.01474:  98%|█████████▊| 246/250 [08:23<00:08,  2.04s/it]\n",
      "Loss: 0.01696:  98%|█████████▊| 246/250 [08:23<00:08,  2.04s/it]\n",
      "Loss: 0.01472:  98%|█████████▊| 246/250 [08:23<00:08,  2.04s/it]\n",
      "Loss: 0.01472:  99%|█████████▉| 247/250 [08:23<00:06,  2.04s/it]\n",
      "Loss: 0.01663:  99%|█████████▉| 247/250 [08:23<00:06,  2.04s/it]\n",
      "Loss: 0.01595:  99%|█████████▉| 247/250 [08:23<00:06,  2.04s/it]\n",
      "Loss: 0.01910:  99%|█████████▉| 247/250 [08:24<00:06,  2.04s/it]\n",
      "Loss: 0.01601:  99%|█████████▉| 247/250 [08:24<00:06,  2.04s/it]\n",
      "Loss: 0.01735:  99%|█████████▉| 247/250 [08:24<00:06,  2.04s/it]\n",
      "Loss: 0.01513:  99%|█████████▉| 247/250 [08:24<00:06,  2.04s/it]\n",
      "Loss: 0.01690:  99%|█████████▉| 247/250 [08:24<00:06,  2.04s/it]\n",
      "Loss: 0.01845:  99%|█████████▉| 247/250 [08:24<00:06,  2.04s/it]\n",
      "Loss: 0.01787:  99%|█████████▉| 247/250 [08:24<00:06,  2.04s/it]\n",
      "Loss: 0.01806:  99%|█████████▉| 247/250 [08:24<00:06,  2.04s/it]\n",
      "Loss: 0.01855:  99%|█████████▉| 247/250 [08:24<00:06,  2.04s/it]\n",
      "Loss: 0.01784:  99%|█████████▉| 247/250 [08:25<00:06,  2.04s/it]\n",
      "Loss: 0.01597:  99%|█████████▉| 247/250 [08:25<00:06,  2.04s/it]\n",
      "Loss: 0.01683:  99%|█████████▉| 247/250 [08:25<00:06,  2.04s/it]\n",
      "Loss: 0.01599:  99%|█████████▉| 247/250 [08:25<00:06,  2.04s/it]\n",
      "Loss: 0.01560:  99%|█████████▉| 247/250 [08:25<00:06,  2.04s/it]\n",
      "Loss: 0.01722:  99%|█████████▉| 247/250 [08:25<00:06,  2.04s/it]\n",
      "Loss: 0.01533:  99%|█████████▉| 247/250 [08:25<00:06,  2.04s/it]\n",
      "Batches: 100%|██████████| 125/125 [00:02<00:00, 61.78it/s]\u001b[A\n",
      "Loss: 0.01533:  99%|█████████▉| 248/250 [08:25<00:04,  2.04s/it]\n",
      "Loss: 0.01694:  99%|█████████▉| 248/250 [08:25<00:04,  2.04s/it]\n",
      "Loss: 0.01746:  99%|█████████▉| 248/250 [08:25<00:04,  2.04s/it]\n",
      "Loss: 0.01836:  99%|█████████▉| 248/250 [08:26<00:04,  2.04s/it]\n",
      "Loss: 0.01605:  99%|█████████▉| 248/250 [08:26<00:04,  2.04s/it]\n",
      "Loss: 0.01666:  99%|█████████▉| 248/250 [08:26<00:04,  2.04s/it]\n",
      "Loss: 0.01543:  99%|█████████▉| 248/250 [08:26<00:04,  2.04s/it]\n",
      "Loss: 0.01608:  99%|█████████▉| 248/250 [08:26<00:04,  2.04s/it]\n",
      "Loss: 0.01659:  99%|█████████▉| 248/250 [08:26<00:04,  2.04s/it]\n",
      "Loss: 0.01782:  99%|█████████▉| 248/250 [08:26<00:04,  2.04s/it]\n",
      "Loss: 0.01761:  99%|█████████▉| 248/250 [08:26<00:04,  2.04s/it]\n",
      "Loss: 0.01656:  99%|█████████▉| 248/250 [08:26<00:04,  2.04s/it]\n",
      "Loss: 0.01683:  99%|█████████▉| 248/250 [08:27<00:04,  2.04s/it]\n",
      "Loss: 0.01746:  99%|█████████▉| 248/250 [08:27<00:04,  2.04s/it]\n",
      "Loss: 0.01514:  99%|█████████▉| 248/250 [08:27<00:04,  2.04s/it]\n",
      "Loss: 0.01750:  99%|█████████▉| 248/250 [08:27<00:04,  2.04s/it]\n",
      "Loss: 0.01829:  99%|█████████▉| 248/250 [08:27<00:04,  2.04s/it]\n",
      "Loss: 0.01869:  99%|█████████▉| 248/250 [08:27<00:04,  2.04s/it]\n",
      "Loss: 0.01592:  99%|█████████▉| 248/250 [08:27<00:04,  2.04s/it]\n",
      "Loss: 0.01528:  99%|█████████▉| 248/250 [08:27<00:04,  2.04s/it]\n",
      "Loss: 0.01528: 100%|█████████▉| 249/250 [08:27<00:02,  2.04s/it]\n",
      "Loss: 0.01736: 100%|█████████▉| 249/250 [08:27<00:02,  2.04s/it]\n",
      "Loss: 0.01706: 100%|█████████▉| 249/250 [08:27<00:02,  2.04s/it]\n",
      "Loss: 0.01910: 100%|█████████▉| 249/250 [08:28<00:02,  2.04s/it]\n",
      "Loss: 0.01924: 100%|█████████▉| 249/250 [08:28<00:02,  2.04s/it]\n",
      "Loss: 0.01794: 100%|█████████▉| 249/250 [08:28<00:02,  2.04s/it]\n",
      "Loss: 0.01822: 100%|█████████▉| 249/250 [08:28<00:02,  2.04s/it]\n",
      "Loss: 0.01718: 100%|█████████▉| 249/250 [08:28<00:02,  2.04s/it]\n",
      "Loss: 0.01820: 100%|█████████▉| 249/250 [08:28<00:02,  2.04s/it]\n",
      "Loss: 0.01665: 100%|█████████▉| 249/250 [08:28<00:02,  2.04s/it]\n",
      "Loss: 0.01869: 100%|█████████▉| 249/250 [08:28<00:02,  2.04s/it]\n",
      "Loss: 0.01724: 100%|█████████▉| 249/250 [08:28<00:02,  2.04s/it]\n",
      "Loss: 0.01549: 100%|█████████▉| 249/250 [08:29<00:02,  2.04s/it]\n",
      "Loss: 0.01977: 100%|█████████▉| 249/250 [08:29<00:02,  2.04s/it]\n",
      "Loss: 0.01762: 100%|█████████▉| 249/250 [08:29<00:02,  2.04s/it]\n",
      "Loss: 0.01652: 100%|█████████▉| 249/250 [08:29<00:02,  2.04s/it]\n",
      "Loss: 0.02035: 100%|█████████▉| 249/250 [08:29<00:02,  2.04s/it]\n",
      "Loss: 0.01532: 100%|█████████▉| 249/250 [08:29<00:02,  2.04s/it]\n",
      "Loss: 0.01736: 100%|█████████▉| 249/250 [08:29<00:02,  2.04s/it]\n",
      "Loss: 0.01467: 100%|█████████▉| 249/250 [08:29<00:02,  2.04s/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, 32, 32)*255):\n",
    "        pil_image = Image.fromarray(image)\n",
    "        pil_image.convert('L').save('CIFAR/imgAEtrain/{}.jpg'.format(i))\n",
    "    for i, image in enumerate(out_test[0].reshape(2000, 32, 32)*255):\n",
    "        pil_image = Image.fromarray(image)\n",
    "        pil_image.convert('L').save('CIFAR/imgAEtest/{}.jpg'.format(i+8000))\n",
    "    result = open('CIFAR/res.txt', 'w')\n",
    "    result.write(str(out_train[1])+'\\n')\n",
    "    result.write(str(out_test[1]))\n",
    "    result.close()\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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
}