lab-09_6_mnist_batchnorm.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Lab 10 MNIST and softmax\n",
    "import torch\n",
    "import torchvision.datasets as dsets\n",
    "import torchvision.transforms as transforms\n",
    "import matplotlib.pylab as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
    "\n",
    "# for reproducibility\n",
    "torch.manual_seed(1)\n",
    "if device == 'cuda':\n",
    "    torch.cuda.manual_seed_all(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# parameters\n",
    "learning_rate = 0.01\n",
    "training_epochs = 10\n",
    "batch_size = 32"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# MNIST dataset\n",
    "mnist_train = dsets.MNIST(root='MNIST_data/',\n",
    "                          train=True,\n",
    "                          transform=transforms.ToTensor(),\n",
    "                          download=True)\n",
    "\n",
    "mnist_test = dsets.MNIST(root='MNIST_data/',\n",
    "                         train=False,\n",
    "                         transform=transforms.ToTensor(),\n",
    "                         download=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# dataset loader\n",
    "train_loader = torch.utils.data.DataLoader(dataset=mnist_train,\n",
    "                                          batch_size=batch_size,\n",
    "                                          shuffle=True,\n",
    "                                          drop_last=True)\n",
    "\n",
    "test_loader = torch.utils.data.DataLoader(dataset=mnist_test,\n",
    "                                          batch_size=batch_size,\n",
    "                                          shuffle=False,\n",
    "                                          drop_last=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# nn layers\n",
    "linear1 = torch.nn.Linear(784, 32, bias=True)\n",
    "linear2 = torch.nn.Linear(32, 32, bias=True)\n",
    "linear3 = torch.nn.Linear(32, 10, bias=True)\n",
    "relu = torch.nn.ReLU()\n",
    "bn1 = torch.nn.BatchNorm1d(32)\n",
    "bn2 = torch.nn.BatchNorm1d(32)\n",
    "\n",
    "nn_linear1 = torch.nn.Linear(784, 32, bias=True)\n",
    "nn_linear2 = torch.nn.Linear(32, 32, bias=True)\n",
    "nn_linear3 = torch.nn.Linear(32, 10, bias=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model\n",
    "bn_model = torch.nn.Sequential(linear1, bn1, relu,\n",
    "                            linear2, bn2, relu,\n",
    "                            linear3).to(device)\n",
    "nn_model = torch.nn.Sequential(nn_linear1, relu,\n",
    "                               nn_linear2, relu,\n",
    "                               nn_linear3).to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define cost/loss & optimizer\n",
    "criterion = torch.nn.CrossEntropyLoss().to(device)    # Softmax is internally computed.\n",
    "bn_optimizer = torch.optim.Adam(bn_model.parameters(), lr=learning_rate)\n",
    "nn_optimizer = torch.optim.Adam(nn_model.parameters(), lr=learning_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Epoch 1-TRAIN] Batchnorm Loss(Acc): bn_loss:0.12363(bn_acc:0.96) vs No Batchnorm Loss(Acc): nn_loss:0.18374(nn_acc:0.95)\n",
      "[Epoch 1-VALID] Batchnorm Loss(Acc): bn_loss:0.12769(bn_acc:0.96) vs No Batchnorm Loss(Acc): nn_loss:0.18346(nn_acc:0.94)\n",
      "\n",
      "[Epoch 2-TRAIN] Batchnorm Loss(Acc): bn_loss:0.10145(bn_acc:0.97) vs No Batchnorm Loss(Acc): nn_loss:0.15138(nn_acc:0.96)\n",
      "[Epoch 2-VALID] Batchnorm Loss(Acc): bn_loss:0.12709(bn_acc:0.96) vs No Batchnorm Loss(Acc): nn_loss:0.18243(nn_acc:0.95)\n",
      "\n",
      "[Epoch 3-TRAIN] Batchnorm Loss(Acc): bn_loss:0.08233(bn_acc:0.97) vs No Batchnorm Loss(Acc): nn_loss:0.14125(nn_acc:0.96)\n",
      "[Epoch 3-VALID] Batchnorm Loss(Acc): bn_loss:0.11058(bn_acc:0.97) vs No Batchnorm Loss(Acc): nn_loss:0.16916(nn_acc:0.95)\n",
      "\n",
      "[Epoch 4-TRAIN] Batchnorm Loss(Acc): bn_loss:0.07632(bn_acc:0.98) vs No Batchnorm Loss(Acc): nn_loss:0.18832(nn_acc:0.94)\n",
      "[Epoch 4-VALID] Batchnorm Loss(Acc): bn_loss:0.10169(bn_acc:0.97) vs No Batchnorm Loss(Acc): nn_loss:0.22995(nn_acc:0.93)\n",
      "\n",
      "[Epoch 5-TRAIN] Batchnorm Loss(Acc): bn_loss:0.07129(bn_acc:0.98) vs No Batchnorm Loss(Acc): nn_loss:0.12536(nn_acc:0.96)\n",
      "[Epoch 5-VALID] Batchnorm Loss(Acc): bn_loss:0.10121(bn_acc:0.97) vs No Batchnorm Loss(Acc): nn_loss:0.17861(nn_acc:0.95)\n",
      "\n",
      "[Epoch 6-TRAIN] Batchnorm Loss(Acc): bn_loss:0.06125(bn_acc:0.98) vs No Batchnorm Loss(Acc): nn_loss:0.13720(nn_acc:0.96)\n",
      "[Epoch 6-VALID] Batchnorm Loss(Acc): bn_loss:0.09136(bn_acc:0.97) vs No Batchnorm Loss(Acc): nn_loss:0.18730(nn_acc:0.95)\n",
      "\n",
      "[Epoch 7-TRAIN] Batchnorm Loss(Acc): bn_loss:0.05966(bn_acc:0.98) vs No Batchnorm Loss(Acc): nn_loss:0.12262(nn_acc:0.97)\n",
      "[Epoch 7-VALID] Batchnorm Loss(Acc): bn_loss:0.09752(bn_acc:0.97) vs No Batchnorm Loss(Acc): nn_loss:0.17871(nn_acc:0.96)\n",
      "\n",
      "[Epoch 8-TRAIN] Batchnorm Loss(Acc): bn_loss:0.05372(bn_acc:0.98) vs No Batchnorm Loss(Acc): nn_loss:0.12606(nn_acc:0.97)\n",
      "[Epoch 8-VALID] Batchnorm Loss(Acc): bn_loss:0.09663(bn_acc:0.97) vs No Batchnorm Loss(Acc): nn_loss:0.17672(nn_acc:0.95)\n",
      "\n",
      "[Epoch 9-TRAIN] Batchnorm Loss(Acc): bn_loss:0.05268(bn_acc:0.98) vs No Batchnorm Loss(Acc): nn_loss:0.10984(nn_acc:0.97)\n",
      "[Epoch 9-VALID] Batchnorm Loss(Acc): bn_loss:0.09833(bn_acc:0.97) vs No Batchnorm Loss(Acc): nn_loss:0.16769(nn_acc:0.96)\n",
      "\n",
      "[Epoch 10-TRAIN] Batchnorm Loss(Acc): bn_loss:0.04630(bn_acc:0.99) vs No Batchnorm Loss(Acc): nn_loss:0.09960(nn_acc:0.97)\n",
      "[Epoch 10-VALID] Batchnorm Loss(Acc): bn_loss:0.09259(bn_acc:0.97) vs No Batchnorm Loss(Acc): nn_loss:0.16750(nn_acc:0.96)\n",
      "\n",
      "Learning finished\n"
     ]
    }
   ],
   "source": [
    "# Save Losses and Accuracies every epoch\n",
    "# We are going to plot them later\n",
    "train_losses = []\n",
    "train_accs = []\n",
    "\n",
    "valid_losses = []\n",
    "valid_accs = []\n",
    "\n",
    "train_total_batch = len(train_loader)\n",
    "test_total_batch = len(test_loader)\n",
    "for epoch in range(training_epochs):\n",
    "    bn_model.train()  # set the model to train mode\n",
    "\n",
    "    for X, Y in train_loader:\n",
    "        # reshape input image into [batch_size by 784]\n",
    "        # label is not one-hot encoded\n",
    "        X = X.view(-1, 28 * 28).to(device)\n",
    "        Y = Y.to(device)\n",
    "\n",
    "        bn_optimizer.zero_grad()\n",
    "        bn_prediction = bn_model(X)\n",
    "        bn_loss = criterion(bn_prediction, Y)\n",
    "        bn_loss.backward()\n",
    "        bn_optimizer.step()\n",
    "\n",
    "        nn_optimizer.zero_grad()\n",
    "        nn_prediction = nn_model(X)\n",
    "        nn_loss = criterion(nn_prediction, Y)\n",
    "        nn_loss.backward()\n",
    "        nn_optimizer.step()\n",
    "\n",
    "    with torch.no_grad():\n",
    "        bn_model.eval()     # set the model to evaluation mode\n",
    "\n",
    "        # Test the model using train sets\n",
    "        bn_loss, nn_loss, bn_acc, nn_acc = 0, 0, 0, 0\n",
    "        for i, (X, Y) in enumerate(train_loader):\n",
    "            X = X.view(-1, 28 * 28).to(device)\n",
    "            Y = Y.to(device)\n",
    "\n",
    "            bn_prediction = bn_model(X)\n",
    "            bn_correct_prediction = torch.argmax(bn_prediction, 1) == Y\n",
    "            bn_loss += criterion(bn_prediction, Y)\n",
    "            bn_acc += bn_correct_prediction.float().mean()\n",
    "\n",
    "            nn_prediction = nn_model(X)\n",
    "            nn_correct_prediction = torch.argmax(nn_prediction, 1) == Y\n",
    "            nn_loss += criterion(nn_prediction, Y)\n",
    "            nn_acc += nn_correct_prediction.float().mean()\n",
    "\n",
    "        bn_loss, nn_loss, bn_acc, nn_acc = bn_loss / train_total_batch, nn_loss / train_total_batch, bn_acc / train_total_batch, nn_acc / train_total_batch\n",
    "\n",
    "        # Save train losses/acc\n",
    "        train_losses.append([bn_loss, nn_loss])\n",
    "        train_accs.append([bn_acc, nn_acc])\n",
    "        print(\n",
    "            '[Epoch %d-TRAIN] Batchnorm Loss(Acc): bn_loss:%.5f(bn_acc:%.2f) vs No Batchnorm Loss(Acc): nn_loss:%.5f(nn_acc:%.2f)' % (\n",
    "            (epoch + 1), bn_loss.item(), bn_acc.item(), nn_loss.item(), nn_acc.item()))\n",
    "        # Test the model using test sets\n",
    "        bn_loss, nn_loss, bn_acc, nn_acc = 0, 0, 0, 0\n",
    "        for i, (X, Y) in enumerate(test_loader):\n",
    "            X = X.view(-1, 28 * 28).to(device)\n",
    "            Y = Y.to(device)\n",
    "\n",
    "            bn_prediction = bn_model(X)\n",
    "            bn_correct_prediction = torch.argmax(bn_prediction, 1) == Y\n",
    "            bn_loss += criterion(bn_prediction, Y)\n",
    "            bn_acc += bn_correct_prediction.float().mean()\n",
    "\n",
    "            nn_prediction = nn_model(X)\n",
    "            nn_correct_prediction = torch.argmax(nn_prediction, 1) == Y\n",
    "            nn_loss += criterion(nn_prediction, Y)\n",
    "            nn_acc += nn_correct_prediction.float().mean()\n",
    "\n",
    "        bn_loss, nn_loss, bn_acc, nn_acc = bn_loss / test_total_batch, nn_loss / test_total_batch, bn_acc / test_total_batch, nn_acc / test_total_batch\n",
    "\n",
    "        # Save valid losses/acc\n",
    "        valid_losses.append([bn_loss, nn_loss])\n",
    "        valid_accs.append([bn_acc, nn_acc])\n",
    "        print(\n",
    "            '[Epoch %d-VALID] Batchnorm Loss(Acc): bn_loss:%.5f(bn_acc:%.2f) vs No Batchnorm Loss(Acc): nn_loss:%.5f(nn_acc:%.2f)' % (\n",
    "                (epoch + 1), bn_loss.item(), bn_acc.item(), nn_loss.item(), nn_acc.item()))\n",
    "        print()\n",
    "\n",
    "print('Learning finished')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_compare(loss_list: list, ylim=None, title=None) -> None:\n",
    "    bn = [i[0] for i in loss_list]\n",
    "    nn = [i[1] for i in loss_list]\n",
    "\n",
    "    plt.figure(figsize=(15, 10))\n",
    "    plt.plot(bn, label='With BN')\n",
    "    plt.plot(nn, label='Without BN')\n",
    "    if ylim:\n",
    "        plt.ylim(ylim)\n",
    "\n",
    "    if title:\n",
    "        plt.title(title)\n",
    "    plt.legend()\n",
    "    plt.grid('on')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xa798b00>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xa798080>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x529aeb8>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x529aef0>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_compare(train_losses, title='Training Loss at Epoch')\n",
    "plot_compare(train_accs, [0, 1.0], title='Training Acc at Epoch')\n",
    "plot_compare(valid_losses, title='Validation Loss at Epoch')\n",
    "plot_compare(valid_accs, [0, 1.0], title='Validation Acc at Epoch')"
   ]
  }
 ],
 "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": "3.5.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}