{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "Deep Learning Models -- A collection of various deep learning architectures, models, and tips for TensorFlow and PyTorch in Jupyter Notebooks.\n", "- Author: Sebastian Raschka\n", "- GitHub Repository: https://github.com/rasbt/deeplearning-models" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Sebastian Raschka \n", "\n", "CPython 3.7.3\n", "IPython 7.6.1\n", "\n", "torch 1.1.0\n" ] } ], "source": [ "%load_ext watermark\n", "%watermark -a 'Sebastian Raschka' -v -p torch" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "- Runs on CPU or GPU (if available)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Model Zoo -- Convolutional Neural Network" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Imports" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "import time\n", "import numpy as np\n", "import torch\n", "import torch.nn.functional as F\n", "from torchvision import datasets\n", "from torchvision import transforms\n", "from torch.utils.data import DataLoader\n", "\n", "\n", "if torch.cuda.is_available():\n", " torch.backends.cudnn.deterministic = True" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Settings and Dataset" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Image batch dimensions: torch.Size([128, 1, 28, 28])\n", "Image label dimensions: torch.Size([128])\n" ] } ], "source": [ "##########################\n", "### SETTINGS\n", "##########################\n", "\n", "# Device\n", "device = torch.device(\"cuda:3\" if torch.cuda.is_available() else \"cpu\")\n", "\n", "# Hyperparameters\n", "random_seed = 1\n", "learning_rate = 0.05\n", "num_epochs = 10\n", "batch_size = 128\n", "\n", "# Architecture\n", "num_classes = 10\n", "\n", "\n", "##########################\n", "### MNIST DATASET\n", "##########################\n", "\n", "# Note transforms.ToTensor() scales input images\n", "# to 0-1 range\n", "train_dataset = datasets.MNIST(root='data', \n", " train=True, \n", " transform=transforms.ToTensor(),\n", " download=True)\n", "\n", "test_dataset = datasets.MNIST(root='data', \n", " train=False, \n", " transform=transforms.ToTensor())\n", "\n", "\n", "train_loader = DataLoader(dataset=train_dataset, \n", " batch_size=batch_size, \n", " shuffle=True)\n", "\n", "test_loader = DataLoader(dataset=test_dataset, \n", " batch_size=batch_size, \n", " shuffle=False)\n", "\n", "# Checking the dataset\n", "for images, labels in train_loader: \n", " print('Image batch dimensions:', images.shape)\n", " print('Image label dimensions:', labels.shape)\n", " break" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Model" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "##########################\n", "### MODEL\n", "##########################\n", "\n", "\n", "class ConvNet(torch.nn.Module):\n", "\n", " def __init__(self, num_classes):\n", " super(ConvNet, self).__init__()\n", " \n", " # calculate same padding:\n", " # (w - k + 2*p)/s + 1 = o\n", " # => p = (s(o-1) - w + k)/2\n", " \n", " # 28x28x1 => 28x28x8\n", " self.conv_1 = torch.nn.Conv2d(in_channels=1,\n", " out_channels=8,\n", " kernel_size=(3, 3),\n", " stride=(1, 1),\n", " padding=1) # (1(28-1) - 28 + 3) / 2 = 1\n", " # 28x28x8 => 14x14x8\n", " self.pool_1 = torch.nn.MaxPool2d(kernel_size=(2, 2),\n", " stride=(2, 2),\n", " padding=0) # (2(14-1) - 28 + 2) = 0 \n", " # 14x14x8 => 14x14x16\n", " self.conv_2 = torch.nn.Conv2d(in_channels=8,\n", " out_channels=16,\n", " kernel_size=(3, 3),\n", " stride=(1, 1),\n", " padding=1) # (1(14-1) - 14 + 3) / 2 = 1 \n", " # 14x14x16 => 7x7x16 \n", " self.pool_2 = torch.nn.MaxPool2d(kernel_size=(2, 2),\n", " stride=(2, 2),\n", " padding=0) # (2(7-1) - 14 + 2) = 0\n", "\n", " self.linear_1 = torch.nn.Linear(7*7*16, num_classes)\n", "\n", " # optionally initialize weights from Gaussian;\n", " # Guassian weight init is not recommended and only for demonstration purposes\n", " for m in self.modules():\n", " if isinstance(m, torch.nn.Conv2d) or isinstance(m, torch.nn.Linear):\n", " m.weight.data.normal_(0.0, 0.01)\n", " m.bias.data.zero_()\n", " if m.bias is not None:\n", " m.bias.detach().zero_()\n", " \n", " \n", " def forward(self, x):\n", " out = self.conv_1(x)\n", " out = F.relu(out)\n", " out = self.pool_1(out)\n", "\n", " out = self.conv_2(out)\n", " out = F.relu(out)\n", " out = self.pool_2(out)\n", " \n", " logits = self.linear_1(out.view(-1, 7*7*16))\n", " probas = F.softmax(logits, dim=1)\n", " return logits, probas\n", "\n", " \n", "torch.manual_seed(random_seed)\n", "model = ConvNet(num_classes=num_classes)\n", "\n", "model = model.to(device)\n", "\n", "optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate) " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Training" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Epoch: 001/010 | Batch 000/469 | Cost: 2.3026\n", "Epoch: 001/010 | Batch 050/469 | Cost: 2.3036\n", "Epoch: 001/010 | Batch 100/469 | Cost: 2.3001\n", "Epoch: 001/010 | Batch 150/469 | Cost: 2.3050\n", "Epoch: 001/010 | Batch 200/469 | Cost: 2.2984\n", "Epoch: 001/010 | Batch 250/469 | Cost: 2.2986\n", "Epoch: 001/010 | Batch 300/469 | Cost: 2.2983\n", "Epoch: 001/010 | Batch 350/469 | Cost: 2.2941\n", "Epoch: 001/010 | Batch 400/469 | Cost: 2.2962\n", "Epoch: 001/010 | Batch 450/469 | Cost: 2.2265\n", "Epoch: 001/010 training accuracy: 65.38%\n", "Time elapsed: 0.24 min\n", "Epoch: 002/010 | Batch 000/469 | Cost: 1.8989\n", "Epoch: 002/010 | Batch 050/469 | Cost: 0.6029\n", "Epoch: 002/010 | Batch 100/469 | Cost: 0.6099\n", "Epoch: 002/010 | Batch 150/469 | Cost: 0.4786\n", "Epoch: 002/010 | Batch 200/469 | Cost: 0.4518\n", "Epoch: 002/010 | Batch 250/469 | Cost: 0.3553\n", "Epoch: 002/010 | Batch 300/469 | Cost: 0.3167\n", "Epoch: 002/010 | Batch 350/469 | Cost: 0.2241\n", "Epoch: 002/010 | Batch 400/469 | Cost: 0.2259\n", "Epoch: 002/010 | Batch 450/469 | Cost: 0.3056\n", "Epoch: 002/010 training accuracy: 93.11%\n", "Time elapsed: 0.47 min\n", "Epoch: 003/010 | Batch 000/469 | Cost: 0.3313\n", "Epoch: 003/010 | Batch 050/469 | Cost: 0.1042\n", "Epoch: 003/010 | Batch 100/469 | Cost: 0.1328\n", "Epoch: 003/010 | Batch 150/469 | Cost: 0.2803\n", "Epoch: 003/010 | Batch 200/469 | Cost: 0.0975\n", "Epoch: 003/010 | Batch 250/469 | Cost: 0.1839\n", "Epoch: 003/010 | Batch 300/469 | Cost: 0.1774\n", "Epoch: 003/010 | Batch 350/469 | Cost: 0.1143\n", "Epoch: 003/010 | Batch 400/469 | Cost: 0.1753\n", "Epoch: 003/010 | Batch 450/469 | Cost: 0.1543\n", "Epoch: 003/010 training accuracy: 95.68%\n", "Time elapsed: 0.70 min\n", "Epoch: 004/010 | Batch 000/469 | Cost: 0.1057\n", "Epoch: 004/010 | Batch 050/469 | Cost: 0.1035\n", "Epoch: 004/010 | Batch 100/469 | Cost: 0.1851\n", "Epoch: 004/010 | Batch 150/469 | Cost: 0.1608\n", "Epoch: 004/010 | Batch 200/469 | Cost: 0.1458\n", "Epoch: 004/010 | Batch 250/469 | Cost: 0.1913\n", "Epoch: 004/010 | Batch 300/469 | Cost: 0.1295\n", "Epoch: 004/010 | Batch 350/469 | Cost: 0.1518\n", "Epoch: 004/010 | Batch 400/469 | Cost: 0.1717\n", "Epoch: 004/010 | Batch 450/469 | Cost: 0.0792\n", "Epoch: 004/010 training accuracy: 96.46%\n", "Time elapsed: 0.93 min\n", "Epoch: 005/010 | Batch 000/469 | Cost: 0.0905\n", "Epoch: 005/010 | Batch 050/469 | Cost: 0.1622\n", "Epoch: 005/010 | Batch 100/469 | Cost: 0.1934\n", "Epoch: 005/010 | Batch 150/469 | Cost: 0.1874\n", "Epoch: 005/010 | Batch 200/469 | Cost: 0.0742\n", "Epoch: 005/010 | Batch 250/469 | Cost: 0.1056\n", "Epoch: 005/010 | Batch 300/469 | Cost: 0.0997\n", "Epoch: 005/010 | Batch 350/469 | Cost: 0.0948\n", "Epoch: 005/010 | Batch 400/469 | Cost: 0.0575\n", "Epoch: 005/010 | Batch 450/469 | Cost: 0.1157\n", "Epoch: 005/010 training accuracy: 96.97%\n", "Time elapsed: 1.16 min\n", "Epoch: 006/010 | Batch 000/469 | Cost: 0.1326\n", "Epoch: 006/010 | Batch 050/469 | Cost: 0.1549\n", "Epoch: 006/010 | Batch 100/469 | Cost: 0.0784\n", "Epoch: 006/010 | Batch 150/469 | Cost: 0.0898\n", "Epoch: 006/010 | Batch 200/469 | Cost: 0.0991\n", "Epoch: 006/010 | Batch 250/469 | Cost: 0.0965\n", "Epoch: 006/010 | Batch 300/469 | Cost: 0.0477\n", "Epoch: 006/010 | Batch 350/469 | Cost: 0.0712\n", "Epoch: 006/010 | Batch 400/469 | Cost: 0.1109\n", "Epoch: 006/010 | Batch 450/469 | Cost: 0.0325\n", "Epoch: 006/010 training accuracy: 97.60%\n", "Time elapsed: 1.39 min\n", "Epoch: 007/010 | Batch 000/469 | Cost: 0.0665\n", "Epoch: 007/010 | Batch 050/469 | Cost: 0.0868\n", "Epoch: 007/010 | Batch 100/469 | Cost: 0.0427\n", "Epoch: 007/010 | Batch 150/469 | Cost: 0.0385\n", "Epoch: 007/010 | Batch 200/469 | Cost: 0.0611\n", "Epoch: 007/010 | Batch 250/469 | Cost: 0.0484\n", "Epoch: 007/010 | Batch 300/469 | Cost: 0.1288\n", "Epoch: 007/010 | Batch 350/469 | Cost: 0.0309\n", "Epoch: 007/010 | Batch 400/469 | Cost: 0.0359\n", "Epoch: 007/010 | Batch 450/469 | Cost: 0.0139\n", "Epoch: 007/010 training accuracy: 97.64%\n", "Time elapsed: 1.62 min\n", "Epoch: 008/010 | Batch 000/469 | Cost: 0.0939\n", "Epoch: 008/010 | Batch 050/469 | Cost: 0.1478\n", "Epoch: 008/010 | Batch 100/469 | Cost: 0.0769\n", "Epoch: 008/010 | Batch 150/469 | Cost: 0.0713\n", "Epoch: 008/010 | Batch 200/469 | Cost: 0.1272\n", "Epoch: 008/010 | Batch 250/469 | Cost: 0.0446\n", "Epoch: 008/010 | Batch 300/469 | Cost: 0.0525\n", "Epoch: 008/010 | Batch 350/469 | Cost: 0.1729\n", "Epoch: 008/010 | Batch 400/469 | Cost: 0.0672\n", "Epoch: 008/010 | Batch 450/469 | Cost: 0.0754\n", "Epoch: 008/010 training accuracy: 96.67%\n", "Time elapsed: 1.85 min\n", "Epoch: 009/010 | Batch 000/469 | Cost: 0.0988\n", "Epoch: 009/010 | Batch 050/469 | Cost: 0.0409\n", "Epoch: 009/010 | Batch 100/469 | Cost: 0.1046\n", "Epoch: 009/010 | Batch 150/469 | Cost: 0.0523\n", "Epoch: 009/010 | Batch 200/469 | Cost: 0.0815\n", "Epoch: 009/010 | Batch 250/469 | Cost: 0.0811\n", "Epoch: 009/010 | Batch 300/469 | Cost: 0.0416\n", "Epoch: 009/010 | Batch 350/469 | Cost: 0.0747\n", "Epoch: 009/010 | Batch 400/469 | Cost: 0.0467\n", "Epoch: 009/010 | Batch 450/469 | Cost: 0.0669\n", "Epoch: 009/010 training accuracy: 97.90%\n", "Time elapsed: 2.08 min\n", "Epoch: 010/010 | Batch 000/469 | Cost: 0.0257\n", "Epoch: 010/010 | Batch 050/469 | Cost: 0.0357\n", "Epoch: 010/010 | Batch 100/469 | Cost: 0.1469\n", "Epoch: 010/010 | Batch 150/469 | Cost: 0.0170\n", "Epoch: 010/010 | Batch 200/469 | Cost: 0.0493\n", "Epoch: 010/010 | Batch 250/469 | Cost: 0.0489\n", "Epoch: 010/010 | Batch 300/469 | Cost: 0.1348\n", "Epoch: 010/010 | Batch 350/469 | Cost: 0.0815\n", "Epoch: 010/010 | Batch 400/469 | Cost: 0.0552\n", "Epoch: 010/010 | Batch 450/469 | Cost: 0.0422\n", "Epoch: 010/010 training accuracy: 97.99%\n", "Time elapsed: 2.31 min\n", "Total Training Time: 2.31 min\n" ] } ], "source": [ "def compute_accuracy(model, data_loader):\n", " correct_pred, num_examples = 0, 0\n", " for features, targets in data_loader:\n", " features = features.to(device)\n", " targets = targets.to(device)\n", " logits, probas = model(features)\n", " _, predicted_labels = torch.max(probas, 1)\n", " num_examples += targets.size(0)\n", " correct_pred += (predicted_labels == targets).sum()\n", " return correct_pred.float()/num_examples * 100\n", " \n", "\n", "start_time = time.time() \n", "for epoch in range(num_epochs):\n", " model = 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", " ### LOGGING\n", " if not batch_idx % 50:\n", " print ('Epoch: %03d/%03d | Batch %03d/%03d | Cost: %.4f' \n", " %(epoch+1, num_epochs, batch_idx, \n", " len(train_loader), cost))\n", " \n", " model = model.eval()\n", " print('Epoch: %03d/%03d training accuracy: %.2f%%' % (\n", " epoch+1, num_epochs, \n", " compute_accuracy(model, train_loader)))\n", "\n", " print('Time elapsed: %.2f min' % ((time.time() - start_time)/60))\n", " \n", "print('Total Training Time: %.2f min' % ((time.time() - start_time)/60))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Evaluation" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Test accuracy: 97.97%\n" ] } ], "source": [ "with torch.set_grad_enabled(False): # save memory during inference\n", " print('Test accuracy: %.2f%%' % (compute_accuracy(model, test_loader)))" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "torch 1.1.0\n", "numpy 1.16.4\n", "torchvision 0.3.0\n", "\n" ] } ], "source": [ "%watermark -iv" ] } ], "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": "3.7.3" }, "toc": { "nav_menu": {}, "number_sections": true, "sideBar": true, "skip_h1_title": false, "title_cell": "Table of Contents", "title_sidebar": "Contents", "toc_cell": false, "toc_position": {}, "toc_section_display": true, "toc_window_display": false } }, "nbformat": 4, "nbformat_minor": 4 }