{ "cells": [ { "cell_type": "markdown", "id": "5f790605", "metadata": {}, "source": [ "# Topic 4: Image classification: MNIST with a small CNN\n", "\n", "So far we have trained models on text. Here we apply the **same training recipe** (model → logits → `F.cross_entropy` → `backward` → optimizer step) to a different domain: classifying 28x28 images of handwritten digits.\n", "Along the way we meet the building blocks of image models: images as 4D tensors, **convolutions**, **pooling**, and the **Adam** optimizer." ] }, { "cell_type": "markdown", "id": "532851bb", "metadata": {}, "source": [ "## 1. Setup, random seed and device\n", "Same as before: fix the seed so everyone gets the same results, and pick the GPU if there is one (everything below also runs fine on a CPU)." ] }, { "cell_type": "code", "execution_count": 1, "id": "2fc4aac2", "metadata": { "execution": { "iopub.execute_input": "2026-10-01T17:09:41.651896Z", "iopub.status.busy": "2026-10-01T17:09:41.651730Z", "iopub.status.idle": "2026-10-01T17:09:49.581410Z", "shell.execute_reply": "2026-10-01T17:09:49.580839Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "torch 2.13.0+cu132 | device: cpu\n" ] } ], "source": [ "import time\n", "import torch\n", "import torch.nn as nn\n", "import torch.nn.functional as F\n", "import torchvision\n", "import matplotlib.pyplot as plt\n", "\n", "torch.manual_seed(1337)\n", "device = \"cuda\" if torch.cuda.is_available() else \"cpu\"\n", "print(\"torch\", torch.__version__, \"| device:\", device)" ] }, { "cell_type": "markdown", "id": "ce19f91c", "metadata": {}, "source": [ "## 2. Load MNIST into tensors\n", "`torchvision.datasets.MNIST` downloads the dataset (60,000 training and 10,000 test images) into `data/`.\n", "We only use it as a downloader: `.data` holds the images as a `uint8` tensor of shape `(N, 28, 28)` with pixel values 0–255, and `.targets` holds the digit labels 0–9." ] }, { "cell_type": "code", "execution_count": 2, "id": "09d36191", "metadata": { "execution": { "iopub.execute_input": "2026-10-01T17:09:49.583647Z", "iopub.status.busy": "2026-10-01T17:09:49.583329Z", "iopub.status.idle": "2026-10-01T17:09:51.501284Z", "shell.execute_reply": "2026-10-01T17:09:51.500758Z" } }, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "\r", " 0%| | 0.00/9.91M [00:00" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, axes = plt.subplots(2, 8, figsize=(8, 2.4))\n", "for i, ax in enumerate(axes.flat):\n", " ax.imshow(x_train[i, 0], cmap=\"gray\")\n", " ax.set_title(int(y_train[i]))\n", " ax.axis(\"off\")\n", "plt.tight_layout()\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "416d1a00", "metadata": {}, "source": [ "## 4. A small convolutional network\n", "A **convolution** (`nn.Conv2d`) slides small learned filters (here 3x3) across the image, so the same few weights detect a pattern (an edge, a curve) anywhere in the image; `padding=1` keeps the height and width unchanged.\n", "**Max pooling** (`F.max_pool2d(x, 2)`) keeps the largest value in each 2x2 block, halving the height and width. After two conv+pool stages we flatten to a vector and finish with ordinary linear layers that output 10 logits, one per digit." ] }, { "cell_type": "code", "execution_count": 6, "id": "373e1704", "metadata": { "execution": { "iopub.execute_input": "2026-10-01T17:09:52.761487Z", "iopub.status.busy": "2026-10-01T17:09:52.761336Z", "iopub.status.idle": "2026-10-01T17:09:52.768963Z", "shell.execute_reply": "2026-10-01T17:09:52.768554Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "SmallCNN(\n", " (conv1): Conv2d(1, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n", " (conv2): Conv2d(16, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n", " (fc1): Linear(in_features=1568, out_features=128, bias=True)\n", " (fc2): Linear(in_features=128, out_features=10, bias=True)\n", ")\n" ] } ], "source": [ "class SmallCNN(nn.Module):\n", " def __init__(self):\n", " super().__init__()\n", " self.conv1 = nn.Conv2d(1, 16, kernel_size=3, padding=1)\n", " self.conv2 = nn.Conv2d(16, 32, kernel_size=3, padding=1)\n", " self.fc1 = nn.Linear(32 * 7 * 7, 128)\n", " self.fc2 = nn.Linear(128, 10)\n", "\n", " def forward(self, x): # x: (N, 1, 28, 28)\n", " x = F.max_pool2d(F.relu(self.conv1(x)), 2) # (N, 16, 14, 14)\n", " x = F.max_pool2d(F.relu(self.conv2(x)), 2) # (N, 32, 7, 7)\n", " x = x.flatten(1) # (N, 32*7*7) = (N, 1568)\n", " x = F.relu(self.fc1(x)) # (N, 128)\n", " return self.fc2(x) # (N, 10) logits\n", "\n", "model = SmallCNN().to(device)\n", "print(model)" ] }, { "cell_type": "markdown", "id": "15083897", "metadata": {}, "source": [ "To check our shape comments, we push a dummy batch of 4 images through the layers one step at a time and print the shape after each." ] }, { "cell_type": "code", "execution_count": 7, "id": "f92cd3e2", "metadata": { "execution": { "iopub.execute_input": "2026-10-01T17:09:52.770763Z", "iopub.status.busy": "2026-10-01T17:09:52.770588Z", "iopub.status.idle": "2026-10-01T17:09:53.141415Z", "shell.execute_reply": "2026-10-01T17:09:53.140932Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "input (4, 1, 28, 28)\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "conv1 + relu (4, 16, 28, 28)\n", "max_pool (4, 16, 14, 14)\n", "conv2 + relu (4, 32, 14, 14)\n", "max_pool (4, 32, 7, 7)\n", "flatten (4, 1568)\n", "fc1 + relu (4, 128)\n", "fc2 (logits) (4, 10)\n" ] } ], "source": [ "x = torch.zeros(4, 1, 28, 28, device=device)\n", "print(\"input \", tuple(x.shape))\n", "x = F.relu(model.conv1(x))\n", "print(\"conv1 + relu \", tuple(x.shape))\n", "x = F.max_pool2d(x, 2)\n", "print(\"max_pool \", tuple(x.shape))\n", "x = F.relu(model.conv2(x))\n", "print(\"conv2 + relu \", tuple(x.shape))\n", "x = F.max_pool2d(x, 2)\n", "print(\"max_pool \", tuple(x.shape))\n", "x = x.flatten(1)\n", "print(\"flatten \", tuple(x.shape))\n", "x = F.relu(model.fc1(x))\n", "print(\"fc1 + relu \", tuple(x.shape))\n", "x = model.fc2(x)\n", "print(\"fc2 (logits) \", tuple(x.shape))" ] }, { "cell_type": "markdown", "id": "06864ca9", "metadata": {}, "source": [ "The convolution layers are tiny because their 3x3 filters are shared across all positions; most of the parameters sit in the first linear layer." ] }, { "cell_type": "code", "execution_count": 8, "id": "a0150efd", "metadata": { "execution": { "iopub.execute_input": "2026-10-01T17:09:53.143305Z", "iopub.status.busy": "2026-10-01T17:09:53.143129Z", "iopub.status.idle": "2026-10-01T17:09:53.146595Z", "shell.execute_reply": "2026-10-01T17:09:53.146170Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "conv1.weight (16, 1, 3, 3) 144\n", "conv1.bias (16,) 16\n", "conv2.weight (32, 16, 3, 3) 4,608\n", "conv2.bias (32,) 32\n", "fc1.weight (128, 1568) 200,704\n", "fc1.bias (128,) 128\n", "fc2.weight (10, 128) 1,280\n", "fc2.bias (10,) 10\n", "total parameters: 206,922\n" ] } ], "source": [ "for name, p in model.named_parameters():\n", " print(f\"{name:12s} {str(tuple(p.shape)):18s} {p.numel():>8,}\")\n", "print(f\"total parameters: {sum(p.numel() for p in model.parameters()):,}\")" ] }, { "cell_type": "markdown", "id": "e984f296", "metadata": {}, "source": [ "## 5. Training\n", "First a small helper to measure **accuracy** (fraction of images whose highest logit is the correct digit). It switches the model to eval mode, turns off gradient tracking, and processes the data in chunks so large sets don't need much memory at once." ] }, { "cell_type": "code", "execution_count": 9, "id": "c42c0538", "metadata": { "execution": { "iopub.execute_input": "2026-10-01T17:09:53.148331Z", "iopub.status.busy": "2026-10-01T17:09:53.148176Z", "iopub.status.idle": "2026-10-01T17:09:54.110483Z", "shell.execute_reply": "2026-10-01T17:09:54.109827Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "accuracy before training: 0.096 (random guessing: 0.100)\n" ] } ], "source": [ "def accuracy(model, x, y, chunk=1000):\n", " model.eval()\n", " correct = 0\n", " with torch.no_grad():\n", " for i in range(0, len(x), chunk):\n", " logits = model(x[i:i + chunk].to(device))\n", " correct += (logits.argmax(dim=1) == y[i:i + chunk].to(device)).sum().item()\n", " model.train()\n", " return correct / len(x)\n", "\n", "print(f\"accuracy before training: {accuracy(model, x_val, y_val):.3f} (random guessing: 0.100)\")" ] }, { "cell_type": "markdown", "id": "de5a8688", "metadata": {}, "source": [ "The training loop is the familiar recipe. Each **epoch** we shuffle the training images with `torch.randperm` and walk through them in minibatches of 64, so every image is seen once per epoch.\n", "New here is **Adam** (`torch.optim.Adam`), a popular optimizer that adapts the step size for each parameter and usually works well with its default learning rate of 1e-3. Two epochs take well under a minute on one CPU core." ] }, { "cell_type": "code", "execution_count": 10, "id": "d20ffc36", "metadata": { "execution": { "iopub.execute_input": "2026-10-01T17:09:54.112544Z", "iopub.status.busy": "2026-10-01T17:09:54.112353Z", "iopub.status.idle": "2026-10-01T17:10:27.351104Z", "shell.execute_reply": "2026-10-01T17:10:27.350444Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "epoch 0 step 200 avg loss (last 200) 0.4426\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "epoch 0 step 400 avg loss (last 200) 0.1375\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "epoch 0 step 600 avg loss (last 200) 0.0934\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "epoch 0 step 800 avg loss (last 200) 0.0789\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "--- epoch 0 done: val accuracy 0.9782 (16s)\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "epoch 1 step 200 avg loss (last 200) 0.0622\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "epoch 1 step 400 avg loss (last 200) 0.0581\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "epoch 1 step 600 avg loss (last 200) 0.0508\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "epoch 1 step 800 avg loss (last 200) 0.0430\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "--- epoch 1 done: val accuracy 0.9740 (33s)\n" ] } ], "source": [ "optimizer = torch.optim.Adam(model.parameters(), lr=1e-3)\n", "batch_size = 64\n", "num_epochs = 2\n", "losses = []\n", "\n", "t0 = time.time()\n", "for epoch in range(num_epochs):\n", " perm = torch.randperm(len(x_train))\n", " running = 0.0\n", " for step, i in enumerate(range(0, len(x_train), batch_size)):\n", " idx = perm[i:i + batch_size]\n", " xb, yb = x_train[idx].to(device), y_train[idx].to(device)\n", "\n", " logits = model(xb)\n", " loss = F.cross_entropy(logits, yb)\n", " optimizer.zero_grad()\n", " loss.backward()\n", " optimizer.step()\n", "\n", " losses.append(loss.item())\n", " running += loss.item()\n", " if (step + 1) % 200 == 0:\n", " print(f\"epoch {epoch} step {step + 1:4d} avg loss (last 200) {running / 200:.4f}\")\n", " running = 0.0\n", " print(f\"--- epoch {epoch} done: val accuracy {accuracy(model, x_val, y_val):.4f} ({time.time() - t0:.0f}s)\")" ] }, { "cell_type": "markdown", "id": "6ce8a74f", "metadata": {}, "source": [ "Plotting the per-minibatch loss shows a fast initial drop followed by slow improvement; the curve is noisy because each point comes from only 64 images." ] }, { "cell_type": "code", "execution_count": 11, "id": "dbdc1f46", "metadata": { "execution": { "iopub.execute_input": "2026-10-01T17:10:27.353305Z", "iopub.status.busy": "2026-10-01T17:10:27.353133Z", "iopub.status.idle": "2026-10-01T17:10:27.670634Z", "shell.execute_reply": "2026-10-01T17:10:27.670113Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plt.figure(figsize=(6, 3))\n", "plt.plot(losses, lw=0.5)\n", "plt.xlabel(\"step\")\n", "plt.ylabel(\"training loss\")\n", "plt.yscale(\"log\")\n", "plt.tight_layout()\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "7d4bcb51", "metadata": {}, "source": [ "## 6. Test accuracy\n", "The test set was never used for training or for choosing anything, so it gives an honest estimate of how well the model works on new digits. Expect about 98–99% after two epochs." ] }, { "cell_type": "code", "execution_count": 12, "id": "30ebb1d6", "metadata": { "execution": { "iopub.execute_input": "2026-10-01T17:10:27.672637Z", "iopub.status.busy": "2026-10-01T17:10:27.672473Z", "iopub.status.idle": "2026-10-01T17:10:40.868677Z", "shell.execute_reply": "2026-10-01T17:10:40.867947Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "train accuracy: 0.9840\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "val accuracy: 0.9740\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "test accuracy: 0.9821\n" ] } ], "source": [ "print(f\"train accuracy: {accuracy(model, x_train, y_train):.4f}\")\n", "print(f\"val accuracy: {accuracy(model, x_val, y_val):.4f}\")\n", "print(f\"test accuracy: {accuracy(model, x_test, y_test):.4f}\")" ] }, { "cell_type": "markdown", "id": "60b48da7", "metadata": {}, "source": [ "## 7. Look at predictions\n", "We predict on the first 16 test images; each title shows the predicted digit (green if correct, red if wrong)." ] }, { "cell_type": "code", "execution_count": 13, "id": "a44b01cf", "metadata": { "execution": { "iopub.execute_input": "2026-10-01T17:10:40.870959Z", "iopub.status.busy": "2026-10-01T17:10:40.870748Z", "iopub.status.idle": "2026-10-01T17:10:41.333503Z", "shell.execute_reply": "2026-10-01T17:10:41.333029Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "model.eval()\n", "with torch.no_grad():\n", " preds = model(x_test[:16].to(device)).argmax(dim=1).cpu()\n", "\n", "fig, axes = plt.subplots(2, 8, figsize=(8, 2.4))\n", "for i, ax in enumerate(axes.flat):\n", " ax.imshow(x_test[i, 0], cmap=\"gray\")\n", " ax.set_title(int(preds[i]), color=\"green\" if preds[i] == y_test[i] else \"red\")\n", " ax.axis(\"off\")\n", "plt.tight_layout()\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "5dd05fd9", "metadata": {}, "source": [ "More instructive are the mistakes. We predict on the whole test set, find the misclassified images, and show a few with their true label and the model's guess; many are genuinely ambiguous or badly written." ] }, { "cell_type": "code", "execution_count": 14, "id": "9a5384c4", "metadata": { "execution": { "iopub.execute_input": "2026-10-01T17:10:41.335319Z", "iopub.status.busy": "2026-10-01T17:10:41.335153Z", "iopub.status.idle": "2026-10-01T17:10:43.990464Z", "shell.execute_reply": "2026-10-01T17:10:43.989989Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "179 of 10000 test images misclassified\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "with torch.no_grad():\n", " all_preds = torch.cat([model(x_test[i:i + 1000].to(device)).argmax(dim=1).cpu()\n", " for i in range(0, len(x_test), 1000)])\n", "wrong = (all_preds != y_test).nonzero().flatten()\n", "print(f\"{len(wrong)} of {len(x_test)} test images misclassified\")\n", "\n", "fig, axes = plt.subplots(2, 8, figsize=(8, 2.6))\n", "for ax, i in zip(axes.flat, wrong[:16]):\n", " ax.imshow(x_test[i, 0], cmap=\"gray\")\n", " ax.set_title(f\"true {int(y_test[i])}\\npred {int(all_preds[i])}\", fontsize=9)\n", " ax.axis(\"off\")\n", "for ax in axes.flat[len(wrong):]:\n", " ax.axis(\"off\")\n", "plt.tight_layout()\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "30e01deb", "metadata": {}, "source": [ "## Takeaways\n", "- The training recipe is identical to the text models: only the data (4D image tensors in `(N, C, H, W)` layout) and the layers (`nn.Conv2d`, max pooling) changed.\n", "- Convolutions share a few small filters across the whole image, so the network reaches about 98% test accuracy after two epochs with only about 200k parameters, most of them in the linear layer.\n", "- Moving to a GPU only requires `model.to(device)` and `xb.to(device)` / `yb.to(device)`; nothing else in the code changes.\n", "- Try training for more epochs, or adding a third conv layer, and see how test accuracy responds.\n", "- Here the whole dataset fits in memory and we batched it by hand. In Topic 5 we use `Dataset`/`DataLoader` for data that doesn't, and save checkpoints so long runs can be resumed." ] } ], "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": "3.13.5" } }, "nbformat": 4, "nbformat_minor": 5 }