wehub-resource-sync
374 lines
11 KiB
Plaintext
374 lines
11 KiB
Plaintext
{
|
|
"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 3.6.8\n",
|
|
"IPython 7.2.0\n",
|
|
"\n",
|
|
"torch 1.0.0\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 -- Softmax Regression"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {},
|
|
"source": [
|
|
"Implementation of softmax regression (multinomial logistic regression)."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {},
|
|
"source": [
|
|
"## Imports"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 2,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"from torchvision import datasets\n",
|
|
"from torchvision import transforms\n",
|
|
"from torch.utils.data import DataLoader\n",
|
|
"import torch.nn.functional as F\n",
|
|
"import torch"
|
|
]
|
|
},
|
|
{
|
|
"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([256, 1, 28, 28])\n",
|
|
"Image label dimensions: torch.Size([256])\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"##########################\n",
|
|
"### SETTINGS\n",
|
|
"##########################\n",
|
|
"\n",
|
|
"# Device\n",
|
|
"device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
|
|
"\n",
|
|
"# Hyperparameters\n",
|
|
"random_seed = 123\n",
|
|
"learning_rate = 0.1\n",
|
|
"num_epochs = 10\n",
|
|
"batch_size = 256\n",
|
|
"\n",
|
|
"# Architecture\n",
|
|
"num_features = 784\n",
|
|
"num_classes = 10\n",
|
|
"\n",
|
|
"\n",
|
|
"##########################\n",
|
|
"### MNIST DATASET\n",
|
|
"##########################\n",
|
|
"\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",
|
|
"\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": "code",
|
|
"execution_count": 4,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"##########################\n",
|
|
"### MODEL\n",
|
|
"##########################\n",
|
|
"\n",
|
|
"class SoftmaxRegression(torch.nn.Module):\n",
|
|
"\n",
|
|
" def __init__(self, num_features, num_classes):\n",
|
|
" super(SoftmaxRegression, self).__init__()\n",
|
|
" self.linear = torch.nn.Linear(num_features, num_classes)\n",
|
|
" \n",
|
|
" self.linear.weight.detach().zero_()\n",
|
|
" self.linear.bias.detach().zero_()\n",
|
|
" \n",
|
|
" def forward(self, x):\n",
|
|
" logits = self.linear(x)\n",
|
|
" probas = F.softmax(logits, dim=1)\n",
|
|
" return logits, probas\n",
|
|
"\n",
|
|
"model = SoftmaxRegression(num_features=num_features,\n",
|
|
" num_classes=num_classes)\n",
|
|
"\n",
|
|
"model.to(device)\n",
|
|
"\n",
|
|
"##########################\n",
|
|
"### COST AND OPTIMIZER\n",
|
|
"##########################\n",
|
|
"\n",
|
|
"optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate) "
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 5,
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Epoch: 001/010 | Batch 000/234 | Cost: 2.3026\n",
|
|
"Epoch: 001/010 | Batch 050/234 | Cost: 0.7941\n",
|
|
"Epoch: 001/010 | Batch 100/234 | Cost: 0.5651\n",
|
|
"Epoch: 001/010 | Batch 150/234 | Cost: 0.4603\n",
|
|
"Epoch: 001/010 | Batch 200/234 | Cost: 0.4822\n",
|
|
"Epoch: 001/010 training accuracy: 88.04%\n",
|
|
"Epoch: 002/010 | Batch 000/234 | Cost: 0.4105\n",
|
|
"Epoch: 002/010 | Batch 050/234 | Cost: 0.4415\n",
|
|
"Epoch: 002/010 | Batch 100/234 | Cost: 0.4367\n",
|
|
"Epoch: 002/010 | Batch 150/234 | Cost: 0.4289\n",
|
|
"Epoch: 002/010 | Batch 200/234 | Cost: 0.3926\n",
|
|
"Epoch: 002/010 training accuracy: 89.37%\n",
|
|
"Epoch: 003/010 | Batch 000/234 | Cost: 0.4112\n",
|
|
"Epoch: 003/010 | Batch 050/234 | Cost: 0.3579\n",
|
|
"Epoch: 003/010 | Batch 100/234 | Cost: 0.3013\n",
|
|
"Epoch: 003/010 | Batch 150/234 | Cost: 0.3258\n",
|
|
"Epoch: 003/010 | Batch 200/234 | Cost: 0.4254\n",
|
|
"Epoch: 003/010 training accuracy: 89.98%\n",
|
|
"Epoch: 004/010 | Batch 000/234 | Cost: 0.3988\n",
|
|
"Epoch: 004/010 | Batch 050/234 | Cost: 0.3690\n",
|
|
"Epoch: 004/010 | Batch 100/234 | Cost: 0.3459\n",
|
|
"Epoch: 004/010 | Batch 150/234 | Cost: 0.4030\n",
|
|
"Epoch: 004/010 | Batch 200/234 | Cost: 0.3240\n",
|
|
"Epoch: 004/010 training accuracy: 90.35%\n",
|
|
"Epoch: 005/010 | Batch 000/234 | Cost: 0.3265\n",
|
|
"Epoch: 005/010 | Batch 050/234 | Cost: 0.3673\n",
|
|
"Epoch: 005/010 | Batch 100/234 | Cost: 0.3085\n",
|
|
"Epoch: 005/010 | Batch 150/234 | Cost: 0.3183\n",
|
|
"Epoch: 005/010 | Batch 200/234 | Cost: 0.3316\n",
|
|
"Epoch: 005/010 training accuracy: 90.64%\n",
|
|
"Epoch: 006/010 | Batch 000/234 | Cost: 0.4518\n",
|
|
"Epoch: 006/010 | Batch 050/234 | Cost: 0.3863\n",
|
|
"Epoch: 006/010 | Batch 100/234 | Cost: 0.3620\n",
|
|
"Epoch: 006/010 | Batch 150/234 | Cost: 0.3733\n",
|
|
"Epoch: 006/010 | Batch 200/234 | Cost: 0.3289\n",
|
|
"Epoch: 006/010 training accuracy: 90.86%\n",
|
|
"Epoch: 007/010 | Batch 000/234 | Cost: 0.3450\n",
|
|
"Epoch: 007/010 | Batch 050/234 | Cost: 0.2289\n",
|
|
"Epoch: 007/010 | Batch 100/234 | Cost: 0.3073\n",
|
|
"Epoch: 007/010 | Batch 150/234 | Cost: 0.2750\n",
|
|
"Epoch: 007/010 | Batch 200/234 | Cost: 0.3456\n",
|
|
"Epoch: 007/010 training accuracy: 91.00%\n",
|
|
"Epoch: 008/010 | Batch 000/234 | Cost: 0.4900\n",
|
|
"Epoch: 008/010 | Batch 050/234 | Cost: 0.3479\n",
|
|
"Epoch: 008/010 | Batch 100/234 | Cost: 0.2343\n",
|
|
"Epoch: 008/010 | Batch 150/234 | Cost: 0.3059\n",
|
|
"Epoch: 008/010 | Batch 200/234 | Cost: 0.3684\n",
|
|
"Epoch: 008/010 training accuracy: 91.22%\n",
|
|
"Epoch: 009/010 | Batch 000/234 | Cost: 0.3762\n",
|
|
"Epoch: 009/010 | Batch 050/234 | Cost: 0.2976\n",
|
|
"Epoch: 009/010 | Batch 100/234 | Cost: 0.2690\n",
|
|
"Epoch: 009/010 | Batch 150/234 | Cost: 0.2610\n",
|
|
"Epoch: 009/010 | Batch 200/234 | Cost: 0.3140\n",
|
|
"Epoch: 009/010 training accuracy: 91.34%\n",
|
|
"Epoch: 010/010 | Batch 000/234 | Cost: 0.2790\n",
|
|
"Epoch: 010/010 | Batch 050/234 | Cost: 0.3070\n",
|
|
"Epoch: 010/010 | Batch 100/234 | Cost: 0.3300\n",
|
|
"Epoch: 010/010 | Batch 150/234 | Cost: 0.2520\n",
|
|
"Epoch: 010/010 | Batch 200/234 | Cost: 0.3301\n",
|
|
"Epoch: 010/010 training accuracy: 91.40%\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"# Manual seed for deterministic data loader\n",
|
|
"torch.manual_seed(random_seed)\n",
|
|
"\n",
|
|
"\n",
|
|
"def compute_accuracy(model, data_loader):\n",
|
|
" correct_pred, num_examples = 0, 0\n",
|
|
" \n",
|
|
" for features, targets in data_loader:\n",
|
|
" features = features.view(-1, 28*28).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",
|
|
" \n",
|
|
" return correct_pred.float() / num_examples * 100\n",
|
|
" \n",
|
|
"\n",
|
|
"for epoch in range(num_epochs):\n",
|
|
" for batch_idx, (features, targets) in enumerate(train_loader):\n",
|
|
" \n",
|
|
" features = features.view(-1, 28*28).to(device)\n",
|
|
" targets = targets.to(device)\n",
|
|
" \n",
|
|
" ### FORWARD AND BACK PROP\n",
|
|
" logits, probas = model(features)\n",
|
|
" \n",
|
|
" # note that the PyTorch implementation of\n",
|
|
" # CrossEntropyLoss works with logits, not\n",
|
|
" # probabilities\n",
|
|
" cost = F.cross_entropy(logits, targets)\n",
|
|
" optimizer.zero_grad()\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_dataset)//batch_size, cost))\n",
|
|
" \n",
|
|
" with torch.set_grad_enabled(False):\n",
|
|
" print('Epoch: %03d/%03d training accuracy: %.2f%%' % (\n",
|
|
" epoch+1, num_epochs, \n",
|
|
" compute_accuracy(model, train_loader)))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 6,
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Test accuracy: 91.77%\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"print('Test accuracy: %.2f%%' % (compute_accuracy(model, test_loader)))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 7,
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"torch 1.0.0\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": "3.7.1"
|
|
},
|
|
"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": 2
|
|
}
|