module05_mnist_conv.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "cdjn6pICZMoM"
   },
   "source": [
    "![alt text](https://engmrk.com/wp-content/uploads/2018/09/LeNet_Original_Image.jpg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "14icu4JgZXqM"
   },
   "source": [
    "![alt text](https://engmrk.com/wp-content/uploads/2018/09/LeNEt_Summary_Table.jpg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "85_HkTjcNZeg"
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "import random\n",
    "import numpy as np\n",
    "\n",
    "random.seed(0)\n",
    "np.random.seed(0)\n",
    "torch.manual_seed(0)\n",
    "torch.cuda.manual_seed(0)\n",
    "torch.backends.cudnn.deterministic = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "kjPVw9UfNZel"
   },
   "outputs": [],
   "source": [
    "import torchvision.datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 279
    },
    "colab_type": "code",
    "id": "buWKcUqGNZeo",
    "outputId": "f2cb4e72-67cd-4262-da9b-228f9e643afd"
   },
   "outputs": [],
   "source": [
    "MNIST_train = torchvision.datasets.MNIST('./', download=True, train=True)\n",
    "MNIST_test = torchvision.datasets.MNIST('./', download=True, train=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 155
    },
    "colab_type": "code",
    "id": "nyrpGOtPNZeq",
    "outputId": "128d4d8d-8ebc-44d7-e8ed-04a92cd7dd40"
   },
   "outputs": [],
   "source": [
    "X_train = MNIST_train.train_data\n",
    "y_train = MNIST_train.train_labels\n",
    "X_test = MNIST_test.test_data\n",
    "y_test = MNIST_test.test_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "f_TmOTWdNZet",
    "outputId": "eb87b5af-7e23-4a1f-bc58-cdfc4774d4aa"
   },
   "outputs": [],
   "source": [
    "len(y_train), len(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 286
    },
    "colab_type": "code",
    "id": "Zd-E4vIpNZe1",
    "outputId": "f72d8415-e392-4af7-9c8f-501eddb5cc9c"
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.imshow(X_train[0, :, :])\n",
    "plt.show()\n",
    "print(y_train[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Tf9LiDGuNZe6"
   },
   "outputs": [],
   "source": [
    "X_train = X_train.unsqueeze(1).float()\n",
    "X_test = X_test.unsqueeze(1).float()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "kxKAhEYHNZe-",
    "outputId": "5b8e9d61-4680-4b6c-8fa1-73e76490d019"
   },
   "outputs": [],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "xhsw9soDODel"
   },
   "outputs": [],
   "source": [
    "class LeNet5(torch.nn.Module):\n",
    "    def __init__(self):\n",
    "        super(LeNet5, self).__init__()\n",
    "        \n",
    "        self.conv1 = torch.nn.Conv2d(\n",
    "            in_channels=1, out_channels=6, kernel_size=5, padding=2)\n",
    "        self.act1  = torch.nn.Tanh()\n",
    "        self.pool1 = torch.nn.AvgPool2d(kernel_size=2, stride=2)\n",
    "       \n",
    "        self.conv2 = torch.nn.Conv2d(\n",
    "            in_channels=6, out_channels=16, kernel_size=5, padding=0)\n",
    "        self.act2  = torch.nn.Tanh()\n",
    "        self.pool2 = torch.nn.AvgPool2d(kernel_size=2, stride=2)\n",
    "        \n",
    "        self.fc1   = torch.nn.Linear(5 * 5 * 16, 120)\n",
    "        self.act3  = torch.nn.Tanh()\n",
    "        \n",
    "        self.fc2   = torch.nn.Linear(120, 84)\n",
    "        self.act4  = torch.nn.Tanh()\n",
    "        \n",
    "        self.fc3   = torch.nn.Linear(84, 10)\n",
    "    \n",
    "    def forward(self, x):\n",
    "        \n",
    "        x = self.conv1(x)\n",
    "        x = self.act1(x)\n",
    "        x = self.pool1(x)\n",
    "        \n",
    "        x = self.conv2(x)\n",
    "        x = self.act2(x)\n",
    "        x = self.pool2(x)\n",
    "        \n",
    "        x = x.view(x.size(0), x.size(1) * x.size(2) * x.size(3))\n",
    "\n",
    "        x = self.fc1(x)\n",
    "        x = self.act3(x)\n",
    "        x = self.fc2(x)\n",
    "        x = self.act4(x)\n",
    "        x = self.fc3(x)\n",
    "        \n",
    "        return x\n",
    "    \n",
    "lenet5 = LeNet5()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "qT-PsO7VNZfB"
   },
   "outputs": [],
   "source": [
    "device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')\n",
    "lenet5 = lenet5.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "rkK0kv1VNZfL"
   },
   "outputs": [],
   "source": [
    "loss = torch.nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(lenet5.parameters(), lr=1.0e-3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 399
    },
    "colab_type": "code",
    "id": "XfA9gqSdNZfQ",
    "outputId": "d849c6a3-e134-4935-813b-b05f63a96ed6"
   },
   "outputs": [],
   "source": [
    "batch_size = 100\n",
    "\n",
    "test_accuracy_history = []\n",
    "test_loss_history = []\n",
    "\n",
    "X_test = X_test.to(device)\n",
    "y_test = y_test.to(device)\n",
    "\n",
    "for epoch in range(10000):\n",
    "    order = np.random.permutation(len(X_train))\n",
    "    for start_index in range(0, len(X_train), batch_size):\n",
    "        optimizer.zero_grad()\n",
    "        \n",
    "        batch_indexes = order[start_index:start_index+batch_size]\n",
    "        \n",
    "        X_batch = X_train[batch_indexes].to(device)\n",
    "        y_batch = y_train[batch_indexes].to(device)\n",
    "        \n",
    "        preds = lenet5.forward(X_batch) \n",
    "        \n",
    "        loss_value = loss(preds, y_batch)\n",
    "        loss_value.backward()\n",
    "        \n",
    "        optimizer.step()\n",
    "        \n",
    "    test_preds = lenet5.forward(X_test)\n",
    "    test_loss_history.append(loss(test_preds, y_test).data.cpu())\n",
    "    \n",
    "    accuracy = (test_preds.argmax(dim=1) == y_test).float().mean().data.cpu()\n",
    "    test_accuracy_history.append(accuracy)\n",
    "    \n",
    "    print(accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "H4hjogO8NZfS"
   },
   "outputs": [],
   "source": [
    "lenet5.forward(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "4FDtagQvNZfU"
   },
   "outputs": [],
   "source": [
    "# plt.plot(test_accuracy_history);\n",
    "plt.plot(test_loss_history);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "vta37dlSNZfZ"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "Lesson 6 Digits Recognition Convolutional Video.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "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"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}