train_wassname.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T02:49:12.797894Z",
     "start_time": "2020-02-02T02:49:12.766850Z"
    }
   },
   "outputs": [],
   "source": [
    "%reload_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T03:23:52.701735Z",
     "start_time": "2020-02-02T03:23:52.666821Z"
    }
   },
   "outputs": [],
   "source": [
    "import argparse\n",
    "\n",
    "import numpy as np\n",
    "import stheno.torch as stheno\n",
    "import torch\n",
    "from tqdm.auto import tqdm\n",
    "from tensorboardX import SummaryWriter\n",
    "import convcnp.data\n",
    "from convcnp.architectures import SimpleConv, UNet\n",
    "from convcnp.cnp import RegressionANP as ANP\n",
    "from convcnp.cnp import RegressionCNP as CNP\n",
    "from convcnp.experiment import (\n",
    "    report_loss,\n",
    "    RunningAverage,\n",
    "    generate_root,\n",
    "    WorkingDirectory,\n",
    "    save_checkpoint\n",
    ")\n",
    "from convcnp.set_conv import ConvCNP\n",
    "from convcnp.utils import device, gaussian_logpdf\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T03:25:42.191028Z",
     "start_time": "2020-02-02T03:25:42.135734Z"
    }
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "def validate(data, model, report_freq=None):\n",
    "    \"\"\"Compute the validation loss.\"\"\"\n",
    "    ravg = RunningAverage()\n",
    "    model.eval()\n",
    "    with torch.no_grad():\n",
    "        for step, task in enumerate(data):\n",
    "            y_mean, y_std = \\\n",
    "                model(task['x_context'], task['y_context'], task['x_target'])\n",
    "            obj = \\\n",
    "                -gaussian_logpdf(task['y_target'], y_mean, y_std,\n",
    "                                 'batched_mean')\n",
    "            ravg.update(obj.item() / data.batch_size, data.batch_size)\n",
    "            if report_freq:\n",
    "                report_loss('Validation', ravg.avg, step, report_freq)\n",
    "    return ravg.avg\n",
    "\n",
    "\n",
    "def train(data, model, opt, report_freq):\n",
    "    \"\"\"Perform a training epoch.\"\"\"\n",
    "    ravg = RunningAverage()\n",
    "    model.train()\n",
    "    for step, task in enumerate(data):\n",
    "        y_mean, y_std = model(task['x_context'], task['y_context'],\n",
    "                              task['x_target'])\n",
    "        obj = -gaussian_logpdf(task['y_target'], y_mean, y_std, 'batched_mean')\n",
    "        obj.backward()\n",
    "        opt.step()\n",
    "        opt.zero_grad()\n",
    "        ravg.update(obj.item() / data.batch_size, data.batch_size)\n",
    "        report_loss('Training', ravg.avg, step, report_freq)\n",
    "    return ravg.avg\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T03:54:27.204014Z",
     "start_time": "2020-02-02T03:54:27.155886Z"
    }
   },
   "outputs": [],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "def to_numpy(x):\n",
    "    \"\"\"Convert a PyTorch tensor to NumPy.\"\"\"\n",
    "    return x.squeeze().detach().cpu().numpy()\n",
    "\n",
    "\n",
    "def plot(gen, model, name):\n",
    "    # Make prediction\n",
    "    task = next(iter(gen))\n",
    "\n",
    "    x_context = task['x_context'][:1]\n",
    "    y_context = task['y_context'][:1]\n",
    "    x_target = task['x_target'][:1]\n",
    "    y_target = task['y_target'][:1]\n",
    "    x_all = task['x'][:1]\n",
    "    y_all = task['y'][:1]\n",
    "\n",
    "\n",
    "    y_mean, y_std = model(x_context, y_context,\n",
    "                                  x_target)\n",
    "\n",
    "    \n",
    "    # sort target for pred & plotting\n",
    "    inds_target = torch.argsort(x_target.squeeze())\n",
    "    x_target = x_target[:, inds_target]\n",
    "    y_target = y_target[:, inds_target]\n",
    "    y_mean = y_mean[:, inds_target]\n",
    "    y_std = y_std[:, inds_target]\n",
    "\n",
    "    # Plot context set.\n",
    "    plt.scatter(to_numpy(x_target), to_numpy(y_target), label='Target Set', s=20, color='grey', alpha=0.5)\n",
    "    plt.scatter(to_numpy(x_context), to_numpy(y_context), label='Context Set', color='black')    \n",
    "    plt.plot(to_numpy(x_all), to_numpy(y_all), '--',  label='Sampled function', color='gray', alpha=0.9)\n",
    "\n",
    "    # Plot model predictions.\n",
    "    plt.plot(to_numpy(x_target), to_numpy(y_mean), label='%s Output' % name, color='blue')\n",
    "    plt.fill_between(to_numpy(x_target),\n",
    "                     to_numpy(y_mean + 1 * y_std),\n",
    "                     to_numpy(y_mean - 1 * y_std),\n",
    "                     color='tab:blue', alpha=0.2)\n",
    "    plt.ylim(-3., 3)\n",
    "    plt.axis('off')\n",
    "    plt.legend()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T02:48:03.075141Z",
     "start_time": "2020-02-02T02:48:02.873326Z"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T03:28:13.738782Z",
     "start_time": "2020-02-02T03:28:13.684405Z"
    }
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "# Parse arguments given to the script.\n",
    "parser = argparse.ArgumentParser()\n",
    "parser.add_argument('data',\n",
    "                    choices=['eq',\n",
    "                             'matern',\n",
    "                             'noisy-mixture',\n",
    "                             'weakly-periodic',\n",
    "                             'sawtooth'],\n",
    "                    help='Data set to train the CNP on. ')\n",
    "parser.add_argument('model',\n",
    "                    choices=['convcnp', 'convcnpxl', 'cnp', 'anp'],\n",
    "                    help='Choice of model. ')\n",
    "parser.add_argument('--root',\n",
    "                    help='Experiment root, which is the directory from which '\n",
    "                         'the experiment will run. If it is not given, '\n",
    "                         'a directory will be automatically created.')\n",
    "parser.add_argument('--train',\n",
    "                    action='store_true',\n",
    "                    help='Perform training. If this is not specified, '\n",
    "                         'the model will be attempted to be loaded from the '\n",
    "                         'experiment root.')\n",
    "parser.add_argument('--epochs',\n",
    "                    default=100,\n",
    "                    type=int,\n",
    "                    help='Number of epochs to train for.')\n",
    "parser.add_argument('--learning_rate',\n",
    "                    default=1e-3,\n",
    "                    type=float,\n",
    "                    help='Learning rate.')\n",
    "parser.add_argument('--weight_decay',\n",
    "                    default=1e-5,\n",
    "                    type=float,\n",
    "                    help='Weight decay.')\n",
    "args = parser.parse_args(['noisy-mixture', 'convcnp', '--train', '--epochs','500'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T03:28:14.379412Z",
     "start_time": "2020-02-02T03:28:14.292491Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Root: _experiments/2020-02-02_11-28-14_convcnpxl-noisy-mixture\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# Load working directory.\n",
    "if args.root:\n",
    "    wd = WorkingDirectory(root=args.root)\n",
    "else:\n",
    "    experiment_name = f'{args.model}-{args.data}'\n",
    "    wd = WorkingDirectory(root=generate_root(experiment_name))\n",
    "\n",
    "# Load data generator.\n",
    "if args.data == 'sawtooth':\n",
    "    gen = convcnp.data.SawtoothGenerator()\n",
    "    gen_val = convcnp.data.SawtoothGenerator(num_tasks=60)\n",
    "    gen_test = convcnp.data.SawtoothGenerator(num_tasks=2048)\n",
    "else:\n",
    "    if args.data == 'eq':\n",
    "        kernel = stheno.EQ().stretch(0.25)\n",
    "    elif args.data == 'matern':\n",
    "        kernel = stheno.Matern52().stretch(0.25)\n",
    "    elif args.data == 'noisy-mixture':\n",
    "        kernel = stheno.EQ().stretch(1.) + \\\n",
    "                 stheno.EQ().stretch(.25) + \\\n",
    "                 0.001 * stheno.Delta()\n",
    "    elif args.data == 'weakly-periodic':\n",
    "        kernel = stheno.EQ().stretch(0.5) * stheno.EQ().periodic(period=0.25)\n",
    "    else:\n",
    "        raise ValueError(f'Unknown data \"{args.data}\".')\n",
    "\n",
    "    gen = convcnp.data.GPGenerator(kernel=kernel)\n",
    "    gen_val = convcnp.data.GPGenerator(kernel=kernel, num_tasks=60)\n",
    "    gen_test = convcnp.data.GPGenerator(kernel=kernel, num_tasks=2048)\n",
    "\n",
    "# Load model.\n",
    "if args.model == 'convcnp':\n",
    "    model = ConvCNP(learn_length_scale=True,\n",
    "                    points_per_unit=64,\n",
    "                    architecture=SimpleConv())\n",
    "elif args.model == 'convcnpxl':\n",
    "    model = ConvCNP(learn_length_scale=True,\n",
    "                    points_per_unit=64,\n",
    "                    architecture=UNet())\n",
    "elif args.model == 'cnp':\n",
    "    model = CNP(latent_dim=128)\n",
    "elif args.model == 'anp':\n",
    "    model = ANP(latent_dim=128)\n",
    "else:\n",
    "    raise ValueError(f'Unknown model {args.model}.')\n",
    "\n",
    "model.to(device)\n",
    "\n",
    "# Perform training.\n",
    "opt = torch.optim.Adam(model.parameters(),\n",
    "                       args.learning_rate,\n",
    "                       weight_decay=args.weight_decay)\n",
    "\n",
    "writer = SummaryWriter(wd.root)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T03:26:01.715250Z",
     "start_time": "2020-02-02T03:26:01.669576Z"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "start_time": "2020-02-02T03:55:42.800Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "65b49d06074742d39f70daecac26f781",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=500.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch: 1/500\n",
      "Training            1: 2.471e+00\n",
      "Training        epoch: 2.447e+00\n",
      "Validation          1: 4.781e-01\n",
      "Validation      epoch: 2.845e+00\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch: 2/500\n",
      "Training            1: 4.406e+00\n",
      "Training        epoch: 2.534e+00\n",
      "Validation          1: 3.291e+00\n",
      "Validation      epoch: 2.806e+00\n",
      "\n",
      "Epoch: 3/500\n",
      "Training            1: 3.401e+00\n",
      "Training        epoch: 2.557e+00\n",
      "Validation          1: 1.292e+00\n",
      "Validation      epoch: 2.772e+00\n",
      "\n",
      "Epoch: 4/500\n",
      "Training            1: 2.842e+00\n",
      "Training        epoch: 2.576e+00\n",
      "Validation          1: 3.556e+00\n",
      "Validation      epoch: 2.391e+00\n",
      "\n",
      "Epoch: 5/500\n",
      "Training            1: 7.803e-01\n",
      "Training        epoch: 2.415e+00\n",
      "Validation          1: 3.174e+00\n",
      "Validation      epoch: 2.657e+00\n",
      "\n",
      "Epoch: 6/500\n",
      "Training            1: 3.948e+00\n",
      "Training        epoch: 2.524e+00\n",
      "Validation          1: 2.069e+00\n",
      "Validation      epoch: 2.439e+00\n",
      "\n",
      "Epoch: 7/500\n",
      "Training            1: 3.960e+00\n",
      "Training        epoch: 2.580e+00\n",
      "Validation          1: 3.747e+00\n",
      "Validation      epoch: 2.523e+00\n",
      "\n",
      "Epoch: 8/500\n",
      "Training            1: 4.327e+00\n",
      "Training        epoch: 2.519e+00\n",
      "Validation          1: 3.717e+00\n",
      "Validation      epoch: 2.580e+00\n",
      "\n",
      "Epoch: 9/500\n",
      "Training            1: 1.702e+00\n",
      "Training        epoch: 2.399e+00\n",
      "Validation          1: 4.281e+00\n",
      "Validation      epoch: 2.575e+00\n",
      "\n",
      "Epoch: 10/500\n",
      "Training            1: 1.265e+00\n",
      "Training        epoch: 2.395e+00\n",
      "Validation          1: 1.077e+00\n",
      "Validation      epoch: 2.627e+00\n",
      "\n",
      "Epoch: 11/500\n",
      "Training            1: 6.849e-01\n",
      "Training        epoch: 2.323e+00\n",
      "Validation          1: 2.813e+00\n",
      "Validation      epoch: 2.493e+00\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch: 12/500\n",
      "Training            1: 5.173e+00\n",
      "Training        epoch: 2.403e+00\n",
      "Validation          1: 7.745e-01\n",
      "Validation      epoch: 2.225e+00\n",
      "\n",
      "Epoch: 13/500\n",
      "Training            1: 3.020e+00\n",
      "Training        epoch: 2.650e+00\n",
      "Validation          1: 3.289e-01\n",
      "Validation      epoch: 2.423e+00\n",
      "\n",
      "Epoch: 14/500\n",
      "Training            1: 5.349e-01\n",
      "Training        epoch: 2.416e+00\n",
      "Validation          1: 2.234e+00\n",
      "Validation      epoch: 2.538e+00\n",
      "\n",
      "Epoch: 15/500\n",
      "Training            1: 3.466e+00\n",
      "Training        epoch: 2.580e+00\n",
      "Validation          1: 2.483e+00\n",
      "Validation      epoch: 2.682e+00\n",
      "\n",
      "Epoch: 16/500\n",
      "Training            1: 4.923e+00\n",
      "Training        epoch: 2.609e+00\n",
      "Validation          1: 1.689e+00\n",
      "Validation      epoch: 2.575e+00\n",
      "\n",
      "Epoch: 17/500\n",
      "Training            1: 1.698e+00\n",
      "Training        epoch: 2.493e+00\n",
      "Validation          1: 3.137e+00\n",
      "Validation      epoch: 2.622e+00\n",
      "\n",
      "Epoch: 18/500\n",
      "Training            1: 2.686e-01\n",
      "Training        epoch: 2.492e+00\n",
      "Validation          1: 2.958e+00\n",
      "Validation      epoch: 2.281e+00\n",
      "\n",
      "Epoch: 19/500\n",
      "Training            1: 4.282e+00\n",
      "Training        epoch: 2.546e+00\n",
      "Validation          1: 1.234e+00\n",
      "Validation      epoch: 2.389e+00\n",
      "\n",
      "Epoch: 20/500\n",
      "Training            1: 7.650e-01\n",
      "Training        epoch: 2.558e+00\n",
      "Validation          1: 1.004e+00\n",
      "Validation      epoch: 2.387e+00\n",
      "\n",
      "Epoch: 21/500\n",
      "Training            1: 3.831e+00\n",
      "Training        epoch: 2.468e+00\n",
      "Validation          1: 8.353e-01\n",
      "Validation      epoch: 2.272e+00\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch: 22/500\n",
      "Training            1: 1.123e+00\n",
      "Training        epoch: 2.502e+00\n",
      "Validation          1: 2.745e+00\n",
      "Validation      epoch: 2.290e+00\n",
      "\n",
      "Epoch: 23/500\n",
      "Training            1: 2.128e-01\n",
      "Training        epoch: 2.424e+00\n",
      "Validation          1: 3.168e+00\n",
      "Validation      epoch: 2.628e+00\n",
      "\n",
      "Epoch: 24/500\n",
      "Training            1: 1.805e+00\n",
      "Training        epoch: 2.442e+00\n",
      "Validation          1: 1.958e+00\n",
      "Validation      epoch: 2.457e+00\n",
      "\n",
      "Epoch: 25/500\n",
      "Training            1: 2.554e+00\n",
      "Training        epoch: 2.484e+00\n",
      "Validation          1: 3.798e+00\n",
      "Validation      epoch: 2.488e+00\n",
      "\n",
      "Epoch: 26/500\n",
      "Training            1: 4.084e+00\n",
      "Training        epoch: 2.485e+00\n",
      "Validation          1: 1.843e-01\n",
      "Validation      epoch: 2.535e+00\n",
      "\n",
      "Epoch: 27/500\n",
      "Training            1: 2.089e+00\n",
      "Training        epoch: 2.388e+00\n",
      "Validation          1: 3.292e-01\n",
      "Validation      epoch: 2.691e+00\n",
      "\n",
      "Epoch: 28/500\n",
      "Training            1: 1.656e+00\n",
      "Training        epoch: 2.399e+00\n",
      "Validation          1: 4.152e+00\n",
      "Validation      epoch: 2.784e+00\n",
      "\n",
      "Epoch: 29/500\n",
      "Training            1: 5.173e-01\n",
      "Training        epoch: 2.354e+00\n",
      "Validation          1: 2.565e+00\n",
      "Validation      epoch: 2.618e+00\n",
      "\n",
      "Epoch: 30/500\n",
      "Training            1: 2.964e+00\n",
      "Training        epoch: 2.473e+00\n",
      "Validation          1: 3.922e+00\n",
      "Validation      epoch: 2.649e+00\n",
      "\n",
      "Epoch: 31/500\n",
      "Training            1: 9.988e-01\n",
      "Training        epoch: 2.479e+00\n",
      "Validation          1: 3.823e-01\n",
      "Validation      epoch: 2.507e+00\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch: 32/500\n",
      "Training            1: 1.034e+00\n",
      "Training        epoch: 2.427e+00\n",
      "Validation          1: 8.345e-01\n",
      "Validation      epoch: 2.201e+00\n",
      "\n",
      "Epoch: 33/500\n",
      "Training            1: 1.693e+00\n",
      "Training        epoch: 2.556e+00\n",
      "Validation          1: 3.479e+00\n",
      "Validation      epoch: 2.502e+00\n",
      "\n",
      "Epoch: 34/500\n",
      "Training            1: 1.931e+00\n",
      "Training        epoch: 2.592e+00\n",
      "Validation          1: 5.081e+00\n",
      "Validation      epoch: 2.866e+00\n",
      "\n",
      "Epoch: 35/500\n",
      "Training            1: 1.237e-01\n",
      "Training        epoch: 2.545e+00\n",
      "Validation          1: 2.187e+00\n",
      "Validation      epoch: 2.419e+00\n",
      "\n",
      "Epoch: 36/500\n",
      "Training            1: 1.983e+00\n",
      "Training        epoch: 2.428e+00\n",
      "Validation          1: 1.600e+00\n",
      "Validation      epoch: 2.558e+00\n",
      "\n",
      "Epoch: 37/500\n",
      "Training            1: 1.917e+00\n",
      "Training        epoch: 2.579e+00\n",
      "Validation          1: 3.300e+00\n",
      "Validation      epoch: 2.541e+00\n",
      "\n",
      "Epoch: 38/500\n",
      "Training            1: 4.043e+00\n",
      "Training        epoch: 2.519e+00\n",
      "Validation          1: 2.558e+00\n",
      "Validation      epoch: 2.791e+00\n",
      "\n",
      "Epoch: 39/500\n",
      "Training            1: 2.356e+00\n",
      "Training        epoch: 2.352e+00\n",
      "Validation          1: 1.823e+00\n",
      "Validation      epoch: 2.755e+00\n",
      "\n",
      "Epoch: 40/500\n",
      "Training            1: 1.287e+00\n",
      "Training        epoch: 2.433e+00\n",
      "Validation          1: 3.140e+00\n",
      "Validation      epoch: 2.285e+00\n",
      "\n",
      "Epoch: 41/500\n",
      "Training            1: 2.624e+00\n",
      "Training        epoch: 2.522e+00\n",
      "Validation          1: 1.471e+00\n",
      "Validation      epoch: 2.561e+00\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch: 42/500\n",
      "Training            1: 4.030e+00\n",
      "Training        epoch: 2.304e+00\n",
      "Validation          1: 3.739e+00\n",
      "Validation      epoch: 2.526e+00\n",
      "\n",
      "Epoch: 43/500\n",
      "Training            1: 4.031e+00\n",
      "Training        epoch: 2.575e+00\n",
      "Validation          1: 2.848e+00\n",
      "Validation      epoch: 2.360e+00\n",
      "\n",
      "Epoch: 44/500\n",
      "Training            1: 3.065e+00\n",
      "Training        epoch: 2.480e+00\n",
      "Validation          1: 1.181e+00\n",
      "Validation      epoch: 2.729e+00\n",
      "\n",
      "Epoch: 45/500\n",
      "Training            1: 4.170e+00\n",
      "Training        epoch: 2.617e+00\n",
      "Validation          1: 4.325e+00\n",
      "Validation      epoch: 2.463e+00\n",
      "\n",
      "Epoch: 46/500\n",
      "Training            1: 1.987e+00\n"
     ]
    }
   ],
   "source": [
    "\n",
    "plot(gen, model, args.model)\n",
    "\n",
    "if args.train:\n",
    "    # Run the training loop, maintaining the best objective value.\n",
    "    best_obj = np.inf\n",
    "    for epoch in tqdm(range(args.epochs)):\n",
    "        print('\\nEpoch: {}/{}'.format(epoch + 1, args.epochs))\n",
    "\n",
    "        # Compute training objective.\n",
    "        train_obj = train(gen, model, opt, report_freq=500)\n",
    "        report_loss('Training', train_obj, 'epoch')\n",
    "        writer.add_scalar('train/loss', val_obj, epoch)\n",
    "\n",
    "        # Compute validation objective.\n",
    "        val_obj = validate(gen_val, model, report_freq=200)\n",
    "        report_loss('Validation', val_obj, 'epoch')\n",
    "        writer.add_scalar('val/loss', val_obj, epoch)\n",
    "        \n",
    "        if epoch%10==0:\n",
    "            plot(gen_val, model, args.model)\n",
    "\n",
    "        # Update the best objective value and checkpoint the model.\n",
    "        is_best = False\n",
    "        if val_obj < best_obj:\n",
    "            best_obj = val_obj\n",
    "            is_best = True\n",
    "        save_checkpoint(wd,\n",
    "                        {'epoch': epoch + 1,\n",
    "                         'state_dict': model.state_dict(),\n",
    "                         'best_acc_top1': best_obj,\n",
    "                         'optimizer': opt.state_dict()},\n",
    "                        is_best=is_best)\n",
    "else:\n",
    "    # Load saved model.\n",
    "    load_dict = torch.load(wd.file('model_best.pth.tar', exists=True))\n",
    "    model.load_state_dict(load_dict['state_dict'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T03:30:44.105743Z",
     "start_time": "2020-02-02T03:28:15.500Z"
    }
   },
   "outputs": [],
   "source": [
    "# Finally, test model on ~2000 tasks.\n",
    "test_obj = validate(gen_test, model)\n",
    "print('Model averages a log-likelihood of %s on unseen tasks.' % -test_obj)\n",
    "with open(wd.file('test_log_likelihood.txt'), 'w') as f:\n",
    "    f.write(str(-test_obj))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Scratch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T03:30:44.108806Z",
     "start_time": "2020-02-02T03:28:16.300Z"
    }
   },
   "outputs": [],
   "source": [
    "gen = convcnp.data.GPGenerator(kernel=kernel)\n",
    "gen_val = convcnp.data.GPGenerator(kernel=kernel, num_tasks=60)\n",
    "gen_test = convcnp.data.GPGenerator(kernel=kernel, num_tasks=2048)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T03:55:35.746820Z",
     "start_time": "2020-02-02T03:55:35.363811Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(gen, model, args.model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T02:39:09.300230Z",
     "start_time": "2020-02-02T02:39:09.296789Z"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T02:41:47.362559Z",
     "start_time": "2020-02-02T02:41:47.264083Z"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-02-02T02:45:43.418734Z",
     "start_time": "2020-02-02T02:45:43.194861Z"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "jup3.7.3",
   "language": "python",
   "name": "jup3.7.3"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": false,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}