AlexNet.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "AlexNet.ipynb",
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Jaisu-1/ResNets-Cifar10/blob/master/AlexNet.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k6A3WjQ9x4uR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import os\n",
        "import time\n",
        "\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "from torch.utils.data import DataLoader\n",
        "from torch.utils.data.dataset import Subset\n",
        "\n",
        "from torchvision import datasets\n",
        "from torchvision import transforms\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "from PIL import Image\n",
        "\n",
        "\n",
        "if torch.cuda.is_available():\n",
        "    torch.backends.cudnn.deterministic = True"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XC3_uOd61D0B",
        "colab_type": "code",
        "outputId": "8c2da100-581c-43e4-a821-fd8462840067",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "device = 'cuda'\n",
        "torch.cuda.is_available()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tIQq9VXWzyoE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Hyperparameters\n",
        "RANDOM_SEED = 1\n",
        "LEARNING_RATE = 0.0001\n",
        "BATCH_SIZE = 256\n",
        "NUM_EPOCHS = 40\n",
        "\n",
        "# Architecture\n",
        "NUM_CLASSES = 10\n",
        "\n",
        "# Other\n",
        "DEVICE = \"cuda:0\""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XFzr8F7Wz1Ka",
        "colab_type": "code",
        "outputId": "6da6605b-940c-4c17-ec1d-4094a1758c73",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "train_indices = torch.arange(0, 48000)\n",
        "valid_indices = torch.arange(48000, 50000)\n",
        "\n",
        "\n",
        "train_transform = transforms.Compose([transforms.Resize((70, 70)),\n",
        "                                      transforms.RandomCrop((64, 64)),\n",
        "                                      transforms.ToTensor()])\n",
        "\n",
        "test_transform = transforms.Compose([transforms.Resize((70, 70)),\n",
        "                                     transforms.CenterCrop((64, 64)),\n",
        "                                     transforms.ToTensor()])\n",
        "\n",
        "train_and_valid = datasets.CIFAR10(root='data', \n",
        "                                   train=True, \n",
        "                                   transform=train_transform,\n",
        "                                   download=True)\n",
        "\n",
        "train_dataset = Subset(train_and_valid, train_indices)\n",
        "valid_dataset = Subset(train_and_valid, valid_indices)\n",
        "test_dataset = datasets.CIFAR10(root='data', \n",
        "                                train=False, \n",
        "                                transform=test_transform,\n",
        "                                download=False)\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "train_loader = DataLoader(dataset=train_dataset, \n",
        "                          batch_size=BATCH_SIZE,\n",
        "                          num_workers=4,\n",
        "                          shuffle=True)\n",
        "\n",
        "valid_loader = DataLoader(dataset=valid_dataset, \n",
        "                          batch_size=BATCH_SIZE,\n",
        "                          num_workers=4,\n",
        "                          shuffle=False)\n",
        "\n",
        "test_loader = DataLoader(dataset=test_dataset, \n",
        "                         batch_size=BATCH_SIZE,\n",
        "                         num_workers=4,\n",
        "                         shuffle=False)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Files already downloaded and verified\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KTc9wf40z7KB",
        "colab_type": "code",
        "outputId": "e663e2cf-4e4b-4e76-a005-ad54ea1de6b1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 229
        }
      },
      "source": [
        "# Checking the dataset\n",
        "print('Training Set:\\n')\n",
        "for images, labels in train_loader:  \n",
        "    print('Image batch dimensions:', images.size())\n",
        "    print('Image label dimensions:', labels.size())\n",
        "    break\n",
        "    \n",
        "# Checking the dataset\n",
        "print('\\nValidation Set:')\n",
        "for images, labels in valid_loader:  \n",
        "    print('Image batch dimensions:', images.size())\n",
        "    print('Image label dimensions:', labels.size())\n",
        "    break\n",
        "\n",
        "# Checking the dataset\n",
        "print('\\nTesting Set:')\n",
        "for images, labels in train_loader:  \n",
        "    print('Image batch dimensions:', images.size())\n",
        "    print('Image label dimensions:', labels.size())\n",
        "    break"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training Set:\n",
            "\n",
            "Image batch dimensions: torch.Size([256, 3, 64, 64])\n",
            "Image label dimensions: torch.Size([256])\n",
            "\n",
            "Validation Set:\n",
            "Image batch dimensions: torch.Size([256, 3, 64, 64])\n",
            "Image label dimensions: torch.Size([256])\n",
            "\n",
            "Testing Set:\n",
            "Image batch dimensions: torch.Size([256, 3, 64, 64])\n",
            "Image label dimensions: torch.Size([256])\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KIZnU1VY0ChC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class AlexNet(nn.Module):\n",
        "\n",
        "    def __init__(self, num_classes):\n",
        "        super(AlexNet, self).__init__()\n",
        "        self.features = nn.Sequential(\n",
        "            nn.Conv2d(3, 64, kernel_size=11, stride=4, padding=2),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.MaxPool2d(kernel_size=3, stride=2),\n",
        "            nn.Conv2d(64, 192, kernel_size=5, padding=2),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.MaxPool2d(kernel_size=3, stride=2),\n",
        "            nn.Conv2d(192, 384, kernel_size=3, padding=1),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.Conv2d(384, 256, kernel_size=3, padding=1),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.Conv2d(256, 256, kernel_size=3, padding=1),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.MaxPool2d(kernel_size=3, stride=2),\n",
        "        )\n",
        "        self.avgpool = nn.AdaptiveAvgPool2d((6, 6))\n",
        "        self.classifier = nn.Sequential(\n",
        "            nn.Dropout(0.5),\n",
        "            nn.Linear(256 * 6 * 6, 4096),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.Dropout(0.5),\n",
        "            nn.Linear(4096, 4096),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.Linear(4096, num_classes)\n",
        "        )\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = self.features(x)\n",
        "        x = self.avgpool(x)\n",
        "        x = x.view(x.size(0), 256 * 6 * 6)\n",
        "        logits = self.classifier(x)\n",
        "        probas = F.softmax(logits, dim=1)\n",
        "        return logits, probas"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KKcwT-0G0KDd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "torch.manual_seed(RANDOM_SEED)\n",
        "\n",
        "model = AlexNet(NUM_CLASSES)\n",
        "model.to(DEVICE)\n",
        "\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=LEARNING_RATE)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Esjdpn2_0uKV",
        "colab_type": "code",
        "outputId": "5c3fd25f-fe94-4ed3-c6b3-ba18427e8c90",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "def compute_acc(model, data_loader, device):\n",
        "    correct_pred, num_examples = 0, 0\n",
        "    model.eval()\n",
        "    for i, (features, targets) in enumerate(data_loader):\n",
        "            \n",
        "        features = features.to(device)\n",
        "        targets = targets.to(device)\n",
        "\n",
        "        logits, probas = model(features)\n",
        "        _, predicted_labels = torch.max(probas, 1)\n",
        "        num_examples += targets.size(0)\n",
        "        assert predicted_labels.size() == targets.size()\n",
        "        correct_pred += (predicted_labels == targets).sum()\n",
        "    return correct_pred.float()/num_examples * 100\n",
        "    \n",
        "\n",
        "start_time = time.time()\n",
        "\n",
        "cost_list = []\n",
        "train_acc_list, valid_acc_list = [], []\n",
        "\n",
        "\n",
        "for epoch in range(NUM_EPOCHS):\n",
        "    \n",
        "    model.train()\n",
        "    for batch_idx, (features, targets) in enumerate(train_loader):\n",
        "        \n",
        "        features = features.to(DEVICE)\n",
        "        targets = targets.to(DEVICE)\n",
        "            \n",
        "        ### FORWARD AND BACK PROP\n",
        "        logits, probas = model(features)\n",
        "        cost = F.cross_entropy(logits, targets)\n",
        "        optimizer.zero_grad()\n",
        "        \n",
        "        cost.backward()\n",
        "        \n",
        "        ### UPDATE MODEL PARAMETERS\n",
        "        optimizer.step()\n",
        "        \n",
        "        #################################################\n",
        "        ### CODE ONLY FOR LOGGING BEYOND THIS POINT\n",
        "        ################################################\n",
        "        cost_list.append(cost.item())\n",
        "        if not batch_idx % 150:\n",
        "            print (f'Epoch: {epoch+1:03d}/{NUM_EPOCHS:03d} | '\n",
        "                   f'Batch {batch_idx:03d}/{len(train_loader):03d} |' \n",
        "                   f' Cost: {cost:.4f}')\n",
        "\n",
        "        \n",
        "\n",
        "    model.eval()\n",
        "    with torch.set_grad_enabled(False): # save memory during inference\n",
        "        \n",
        "        train_acc = compute_acc(model, train_loader, device=DEVICE)\n",
        "        valid_acc = compute_acc(model, valid_loader, device=DEVICE)\n",
        "        \n",
        "        print(f'Epoch: {epoch+1:03d}/{NUM_EPOCHS:03d}\\n'\n",
        "              f'Train ACC: {train_acc:.2f} | Validation ACC: {valid_acc:.2f}')\n",
        "        \n",
        "        train_acc_list.append(train_acc)\n",
        "        valid_acc_list.append(valid_acc)\n",
        "        \n",
        "    elapsed = (time.time() - start_time)/60\n",
        "    print(f'Time elapsed: {elapsed:.2f} min')\n",
        "  \n",
        "elapsed = (time.time() - start_time)/60\n",
        "print(f'Total Training Time: {elapsed:.2f} min')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch: 001/040 | Batch 000/188 | Cost: 2.3028\n",
            "Epoch: 001/040 | Batch 150/188 | Cost: 1.7150\n",
            "Epoch: 001/040\n",
            "Train ACC: 32.38 | Validation ACC: 31.70\n",
            "Time elapsed: 0.51 min\n",
            "Epoch: 002/040 | Batch 000/188 | Cost: 1.7091\n",
            "Epoch: 002/040 | Batch 150/188 | Cost: 1.5775\n",
            "Epoch: 002/040\n",
            "Train ACC: 45.36 | Validation ACC: 45.80\n",
            "Time elapsed: 1.02 min\n",
            "Epoch: 003/040 | Batch 000/188 | Cost: 1.4383\n",
            "Epoch: 003/040 | Batch 150/188 | Cost: 1.4720\n",
            "Epoch: 003/040\n",
            "Train ACC: 52.43 | Validation ACC: 51.95\n",
            "Time elapsed: 1.53 min\n",
            "Epoch: 004/040 | Batch 000/188 | Cost: 1.3928\n",
            "Epoch: 004/040 | Batch 150/188 | Cost: 1.2750\n",
            "Epoch: 004/040\n",
            "Train ACC: 57.55 | Validation ACC: 57.10\n",
            "Time elapsed: 2.04 min\n",
            "Epoch: 005/040 | Batch 000/188 | Cost: 1.0782\n",
            "Epoch: 005/040 | Batch 150/188 | Cost: 1.1286\n",
            "Epoch: 005/040\n",
            "Train ACC: 61.80 | Validation ACC: 61.25\n",
            "Time elapsed: 2.55 min\n",
            "Epoch: 006/040 | Batch 000/188 | Cost: 1.0204\n",
            "Epoch: 006/040 | Batch 150/188 | Cost: 1.0620\n",
            "Epoch: 006/040\n",
            "Train ACC: 62.41 | Validation ACC: 61.75\n",
            "Time elapsed: 3.06 min\n",
            "Epoch: 007/040 | Batch 000/188 | Cost: 1.0431\n",
            "Epoch: 007/040 | Batch 150/188 | Cost: 1.0323\n",
            "Epoch: 007/040\n",
            "Train ACC: 65.89 | Validation ACC: 64.25\n",
            "Time elapsed: 3.57 min\n",
            "Epoch: 008/040 | Batch 000/188 | Cost: 1.0755\n",
            "Epoch: 008/040 | Batch 150/188 | Cost: 0.8886\n",
            "Epoch: 008/040\n",
            "Train ACC: 66.61 | Validation ACC: 66.10\n",
            "Time elapsed: 4.08 min\n",
            "Epoch: 009/040 | Batch 000/188 | Cost: 0.9461\n",
            "Epoch: 009/040 | Batch 150/188 | Cost: 0.7461\n",
            "Epoch: 009/040\n",
            "Train ACC: 70.68 | Validation ACC: 67.20\n",
            "Time elapsed: 4.59 min\n",
            "Epoch: 010/040 | Batch 000/188 | Cost: 0.7602\n",
            "Epoch: 010/040 | Batch 150/188 | Cost: 0.7927\n",
            "Epoch: 010/040\n",
            "Train ACC: 72.13 | Validation ACC: 68.90\n",
            "Time elapsed: 5.09 min\n",
            "Epoch: 011/040 | Batch 000/188 | Cost: 0.7938\n",
            "Epoch: 011/040 | Batch 150/188 | Cost: 0.7701\n",
            "Epoch: 011/040\n",
            "Train ACC: 73.20 | Validation ACC: 69.00\n",
            "Time elapsed: 5.60 min\n",
            "Epoch: 012/040 | Batch 000/188 | Cost: 0.7008\n",
            "Epoch: 012/040 | Batch 150/188 | Cost: 0.9219\n",
            "Epoch: 012/040\n",
            "Train ACC: 75.95 | Validation ACC: 69.70\n",
            "Time elapsed: 6.11 min\n",
            "Epoch: 013/040 | Batch 000/188 | Cost: 0.6396\n",
            "Epoch: 013/040 | Batch 150/188 | Cost: 0.6736\n",
            "Epoch: 013/040\n",
            "Train ACC: 78.10 | Validation ACC: 70.80\n",
            "Time elapsed: 6.62 min\n",
            "Epoch: 014/040 | Batch 000/188 | Cost: 0.6384\n",
            "Epoch: 014/040 | Batch 150/188 | Cost: 0.6315\n",
            "Epoch: 014/040\n",
            "Train ACC: 78.39 | Validation ACC: 71.60\n",
            "Time elapsed: 7.13 min\n",
            "Epoch: 015/040 | Batch 000/188 | Cost: 0.6078\n",
            "Epoch: 015/040 | Batch 150/188 | Cost: 0.8047\n",
            "Epoch: 015/040\n",
            "Train ACC: 79.54 | Validation ACC: 71.10\n",
            "Time elapsed: 7.64 min\n",
            "Epoch: 016/040 | Batch 000/188 | Cost: 0.7112\n",
            "Epoch: 016/040 | Batch 150/188 | Cost: 0.5460\n",
            "Epoch: 016/040\n",
            "Train ACC: 80.79 | Validation ACC: 70.40\n",
            "Time elapsed: 8.15 min\n",
            "Epoch: 017/040 | Batch 000/188 | Cost: 0.5900\n",
            "Epoch: 017/040 | Batch 150/188 | Cost: 0.5732\n",
            "Epoch: 017/040\n",
            "Train ACC: 83.19 | Validation ACC: 72.15\n",
            "Time elapsed: 8.66 min\n",
            "Epoch: 018/040 | Batch 000/188 | Cost: 0.4973\n",
            "Epoch: 018/040 | Batch 150/188 | Cost: 0.4567\n",
            "Epoch: 018/040\n",
            "Train ACC: 82.98 | Validation ACC: 71.90\n",
            "Time elapsed: 9.17 min\n",
            "Epoch: 019/040 | Batch 000/188 | Cost: 0.4819\n",
            "Epoch: 019/040 | Batch 150/188 | Cost: 0.4989\n",
            "Epoch: 019/040\n",
            "Train ACC: 85.27 | Validation ACC: 72.15\n",
            "Time elapsed: 9.68 min\n",
            "Epoch: 020/040 | Batch 000/188 | Cost: 0.4288\n",
            "Epoch: 020/040 | Batch 150/188 | Cost: 0.3886\n",
            "Epoch: 020/040\n",
            "Train ACC: 87.08 | Validation ACC: 74.25\n",
            "Time elapsed: 10.19 min\n",
            "Epoch: 021/040 | Batch 000/188 | Cost: 0.4047\n",
            "Epoch: 021/040 | Batch 150/188 | Cost: 0.5063\n",
            "Epoch: 021/040\n",
            "Train ACC: 87.46 | Validation ACC: 74.00\n",
            "Time elapsed: 10.70 min\n",
            "Epoch: 022/040 | Batch 000/188 | Cost: 0.4049\n",
            "Epoch: 022/040 | Batch 150/188 | Cost: 0.3497\n",
            "Epoch: 022/040\n",
            "Train ACC: 88.62 | Validation ACC: 72.00\n",
            "Time elapsed: 11.21 min\n",
            "Epoch: 023/040 | Batch 000/188 | Cost: 0.3010\n",
            "Epoch: 023/040 | Batch 150/188 | Cost: 0.4002\n",
            "Epoch: 023/040\n",
            "Train ACC: 90.06 | Validation ACC: 73.95\n",
            "Time elapsed: 11.72 min\n",
            "Epoch: 024/040 | Batch 000/188 | Cost: 0.3040\n",
            "Epoch: 024/040 | Batch 150/188 | Cost: 0.4010\n",
            "Epoch: 024/040\n",
            "Train ACC: 90.71 | Validation ACC: 73.20\n",
            "Time elapsed: 12.23 min\n",
            "Epoch: 025/040 | Batch 000/188 | Cost: 0.2049\n",
            "Epoch: 025/040 | Batch 150/188 | Cost: 0.3291\n",
            "Epoch: 025/040\n",
            "Train ACC: 90.84 | Validation ACC: 72.15\n",
            "Time elapsed: 12.73 min\n",
            "Epoch: 026/040 | Batch 000/188 | Cost: 0.2793\n",
            "Epoch: 026/040 | Batch 150/188 | Cost: 0.3879\n",
            "Epoch: 026/040\n",
            "Train ACC: 91.00 | Validation ACC: 73.75\n",
            "Time elapsed: 13.24 min\n",
            "Epoch: 027/040 | Batch 000/188 | Cost: 0.2951\n",
            "Epoch: 027/040 | Batch 150/188 | Cost: 0.2654\n",
            "Epoch: 027/040\n",
            "Train ACC: 91.28 | Validation ACC: 72.35\n",
            "Time elapsed: 13.75 min\n",
            "Epoch: 028/040 | Batch 000/188 | Cost: 0.2679\n",
            "Epoch: 028/040 | Batch 150/188 | Cost: 0.2585\n",
            "Epoch: 028/040\n",
            "Train ACC: 93.46 | Validation ACC: 72.30\n",
            "Time elapsed: 14.26 min\n",
            "Epoch: 029/040 | Batch 000/188 | Cost: 0.1676\n",
            "Epoch: 029/040 | Batch 150/188 | Cost: 0.2184\n",
            "Epoch: 029/040\n",
            "Train ACC: 93.09 | Validation ACC: 72.40\n",
            "Time elapsed: 14.77 min\n",
            "Epoch: 030/040 | Batch 000/188 | Cost: 0.1518\n",
            "Epoch: 030/040 | Batch 150/188 | Cost: 0.1950\n",
            "Epoch: 030/040\n",
            "Train ACC: 93.79 | Validation ACC: 72.95\n",
            "Time elapsed: 15.28 min\n",
            "Epoch: 031/040 | Batch 000/188 | Cost: 0.1990\n",
            "Epoch: 031/040 | Batch 150/188 | Cost: 0.2045\n",
            "Epoch: 031/040\n",
            "Train ACC: 94.94 | Validation ACC: 72.55\n",
            "Time elapsed: 15.79 min\n",
            "Epoch: 032/040 | Batch 000/188 | Cost: 0.1539\n",
            "Epoch: 032/040 | Batch 150/188 | Cost: 0.2123\n",
            "Epoch: 032/040\n",
            "Train ACC: 94.44 | Validation ACC: 72.35\n",
            "Time elapsed: 16.30 min\n",
            "Epoch: 033/040 | Batch 000/188 | Cost: 0.1391\n",
            "Epoch: 033/040 | Batch 150/188 | Cost: 0.1589\n",
            "Epoch: 033/040\n",
            "Train ACC: 93.48 | Validation ACC: 71.95\n",
            "Time elapsed: 16.81 min\n",
            "Epoch: 034/040 | Batch 000/188 | Cost: 0.1543\n",
            "Epoch: 034/040 | Batch 150/188 | Cost: 0.1709\n",
            "Epoch: 034/040\n",
            "Train ACC: 95.12 | Validation ACC: 73.35\n",
            "Time elapsed: 17.32 min\n",
            "Epoch: 035/040 | Batch 000/188 | Cost: 0.1666\n",
            "Epoch: 035/040 | Batch 150/188 | Cost: 0.1523\n",
            "Epoch: 035/040\n",
            "Train ACC: 96.32 | Validation ACC: 73.75\n",
            "Time elapsed: 17.83 min\n",
            "Epoch: 036/040 | Batch 000/188 | Cost: 0.1200\n",
            "Epoch: 036/040 | Batch 150/188 | Cost: 0.1781\n",
            "Epoch: 036/040\n",
            "Train ACC: 93.91 | Validation ACC: 72.70\n",
            "Time elapsed: 18.34 min\n",
            "Epoch: 037/040 | Batch 000/188 | Cost: 0.1772\n",
            "Epoch: 037/040 | Batch 150/188 | Cost: 0.1422\n",
            "Epoch: 037/040\n",
            "Train ACC: 95.74 | Validation ACC: 73.75\n",
            "Time elapsed: 18.85 min\n",
            "Epoch: 038/040 | Batch 000/188 | Cost: 0.1461\n",
            "Epoch: 038/040 | Batch 150/188 | Cost: 0.1433\n",
            "Epoch: 038/040\n",
            "Train ACC: 95.81 | Validation ACC: 73.60\n",
            "Time elapsed: 19.36 min\n",
            "Epoch: 039/040 | Batch 000/188 | Cost: 0.1041\n",
            "Epoch: 039/040 | Batch 150/188 | Cost: 0.0882\n",
            "Epoch: 039/040\n",
            "Train ACC: 96.60 | Validation ACC: 73.15\n",
            "Time elapsed: 19.87 min\n",
            "Epoch: 040/040 | Batch 000/188 | Cost: 0.1099\n",
            "Epoch: 040/040 | Batch 150/188 | Cost: 0.1332\n",
            "Epoch: 040/040\n",
            "Train ACC: 97.23 | Validation ACC: 74.05\n",
            "Time elapsed: 20.38 min\n",
            "Total Training Time: 20.38 min\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7Zgkk0P95ZDJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o4aivUdw5b1h",
        "colab_type": "code",
        "outputId": "213295e3-5dca-49c1-c5f9-b73bc5984b68",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "next(model.parameters()).is_cuda"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wqR8lD0u5dsi",
        "colab_type": "code",
        "outputId": "849a0eb5-b50a-4897-a19f-6f989241a8c6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        }
      },
      "source": [
        "plt.plot(cost_list, label='Minibatch cost')\n",
        "plt.plot(np.convolve(cost_list, \n",
        "                     np.ones(200,)/200, mode='valid'), \n",
        "         label='Running average')\n",
        "\n",
        "plt.ylabel('Cross Entropy')\n",
        "plt.xlabel('Iteration')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0xaMe2SE5rSa",
        "colab_type": "code",
        "outputId": "077050fd-76bd-4597-83ba-a51dcbd3e763",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        }
      },
      "source": [
        "plt.plot(np.arange(1, NUM_EPOCHS+1), train_acc_list, label='Training')\n",
        "plt.plot(np.arange(1, NUM_EPOCHS+1), valid_acc_list, label='Validation')\n",
        "\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H_yLb8buCmIp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7OmB-nj55w5l",
        "colab_type": "code",
        "outputId": "26ea3d32-c524-4ce6-db76-baa7929503bf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "with torch.set_grad_enabled(False):\n",
        "    test_acc = compute_acc(model=model,\n",
        "                           data_loader=test_loader,\n",
        "                           device=DEVICE)\n",
        "    \n",
        "    valid_acc = compute_acc(model=model,\n",
        "                            data_loader=valid_loader,\n",
        "                            device=DEVICE)\n",
        "    \n",
        "\n",
        "print(f'Validation ACC: {valid_acc:.2f}%')\n",
        "print(f'Test ACC: {test_acc:.2f}%')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Validation ACC: 73.10%\n",
            "Test ACC: 73.38%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N1zs8cGs701d",
        "colab_type": "code",
        "outputId": "d45c65c2-6cad-4eab-a261-ab804e36b087",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "print(\"!Done\")"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "!DOne\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9BDEyrmfA_mJ",
        "colab_type": "code",
        "outputId": "ffab47b3-c9f8-4631-a126-39b8f8ca023e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 618
        }
      },
      "source": [
        "from torchsummary import summary\n",
        "summary(model, (3, 64, 364))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "            Conv2d-1           [-1, 64, 15, 90]          23,296\n",
            "              ReLU-2           [-1, 64, 15, 90]               0\n",
            "         MaxPool2d-3            [-1, 64, 7, 44]               0\n",
            "            Conv2d-4           [-1, 192, 7, 44]         307,392\n",
            "              ReLU-5           [-1, 192, 7, 44]               0\n",
            "         MaxPool2d-6           [-1, 192, 3, 21]               0\n",
            "            Conv2d-7           [-1, 384, 3, 21]         663,936\n",
            "              ReLU-8           [-1, 384, 3, 21]               0\n",
            "            Conv2d-9           [-1, 256, 3, 21]         884,992\n",
            "             ReLU-10           [-1, 256, 3, 21]               0\n",
            "           Conv2d-11           [-1, 256, 3, 21]         590,080\n",
            "             ReLU-12           [-1, 256, 3, 21]               0\n",
            "        MaxPool2d-13           [-1, 256, 1, 10]               0\n",
            "AdaptiveAvgPool2d-14            [-1, 256, 6, 6]               0\n",
            "          Dropout-15                 [-1, 9216]               0\n",
            "           Linear-16                 [-1, 4096]      37,752,832\n",
            "             ReLU-17                 [-1, 4096]               0\n",
            "          Dropout-18                 [-1, 4096]               0\n",
            "           Linear-19                 [-1, 4096]      16,781,312\n",
            "             ReLU-20                 [-1, 4096]               0\n",
            "           Linear-21                   [-1, 10]          40,970\n",
            "================================================================\n",
            "Total params: 57,044,810\n",
            "Trainable params: 57,044,810\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 0.27\n",
            "Forward/backward pass size (MB): 3.64\n",
            "Params size (MB): 217.61\n",
            "Estimated Total Size (MB): 221.52\n",
            "----------------------------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XpHH_eaJC01x",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4AgCd2PQDDfQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_CVB5VEzDSFJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H3RQTUFZB6MH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JdT18TLhCI5U",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4J_J2dROCXis",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tgKoY1E2BOPp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PqaVFKscBc53",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}