{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Image Classification with CNNs\n", "\n", "In this tutorial, we'll build a complete image classification system using Convolutional Neural Networks (CNNs) in BrainState.\n", "\n", "## Learning Objectives\n", "\n", "By the end of this tutorial, you will be able to:\n", "- Build CNN architectures for image classification\n", "- Load and preprocess image datasets (MNIST-like)\n", "- Implement complete training loops\n", "- Evaluate model performance\n", "- Visualize results and predictions\n", "- Apply data augmentation\n", "- Monitor training progress\n", "\n", "## What We'll Build\n", "\n", "We'll create:\n", "- A CNN architecture from scratch\n", "- A training pipeline with validation\n", "- Evaluation metrics and visualization\n", "- A complete end-to-end workflow" ] }, { "cell_type": "code", "execution_count": 1, "metadata": { "ExecuteTime": { "end_time": "2025-10-11T10:06:48.994047Z", "start_time": "2025-10-11T10:06:47.316860Z" }, "execution": { "iopub.execute_input": "2026-05-30T17:32:05.575201Z", "iopub.status.busy": "2026-05-30T17:32:05.574974Z", "iopub.status.idle": "2026-05-30T17:32:10.106806Z", "shell.execute_reply": "2026-05-30T17:32:10.106010Z" } }, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "An NVIDIA GPU may be present on this machine, but a CUDA-enabled jaxlib is not installed. Falling back to cpu.\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "JAX devices: [CpuDevice(id=0)]\n" ] } ], "source": [ "import brainstate\n", "import braintools\n", "import jax\n", "import jax.numpy as jnp\n", "import matplotlib.pyplot as plt\n", "import numpy as np\n", "import time\n", "\n", "# Set random seed for reproducibility\n", "brainstate.random.seed(42)\n", "\n", "print(f\"JAX devices: {jax.devices()}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 1. Dataset Preparation\n", "\n", "We'll create synthetic MNIST-like data for demonstration." ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "ExecuteTime": { "end_time": "2025-10-11T10:07:15.863644Z", "start_time": "2025-10-11T10:06:51.731643Z" }, "execution": { "iopub.execute_input": "2026-05-30T17:32:10.109462Z", "iopub.status.busy": "2026-05-30T17:32:10.108991Z", "iopub.status.idle": "2026-05-30T17:36:28.683666Z", "shell.execute_reply": "2026-05-30T17:36:28.682901Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Generating synthetic dataset...\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Training set: (5000, 28, 28, 1), (5000,)\n", "Validation set: (500, 28, 28, 1), (500,)\n", "Test set: (1000, 28, 28, 1), (1000,)\n" ] } ], "source": [ "def generate_synthetic_mnist(n_samples: int = 1000, img_size: int = 28, n_classes: int = 10):\n", " \"\"\"Generate synthetic MNIST-like dataset.\n", " \n", " Args:\n", " n_samples: Number of samples to generate\n", " img_size: Image size (height and width)\n", " n_classes: Number of classes\n", " \n", " Returns:\n", " images: Array of shape (n_samples, img_size, img_size, 1)\n", " labels: Array of shape (n_samples,)\n", " \"\"\"\n", " # Generate random images with patterns\n", " images = brainstate.random.randn(n_samples, img_size, img_size, 1) * 0.4\n", " labels = brainstate.random.randint(0, n_classes, (n_samples,))\n", " \n", " # Add class-specific patterns\n", " for i in range(n_samples):\n", " label = int(labels[i])\n", " \n", " \n", " if label == 0: # circle\n", " center_x, center_y = img_size//2, img_size//2\n", " radius = 8\n", " for x in range(img_size):\n", " for y in range(img_size):\n", " dist = (x-center_x)**2 + (y-center_y)**2\n", " if radius**2 - 3 <= dist <= radius**2 + 3:\n", " images = images.at[i, x, y, 0].set(0.9)\n", " \n", " elif label == 1: # Vertical line + bottom small horizontal line (similar to 7)\n", " center = img_size//2\n", " images = images.at[i, center-1:center+1, 5:img_size-5, 0].set(0.9)\n", " images = images.at[i, img_size-6:img_size-4, center-3:center+3, 0].set(0.9)\n", " \n", " elif label == 2: \n", " images = images.at[i, 5, 5:img_size-5, 0].set(0.9)\n", " images = images.at[i, 6, 5:img_size-5, 0].set(0.9)\n", " for j in range(10):\n", " images = images.at[i, 6+j, img_size-6-j, 0].set(0.9)\n", " images = images.at[i, 7+j, img_size-6-j, 0].set(0.9)\n", " images = images.at[i, img_size-6, 5:img_size-5, 0].set(0.9)\n", " images = images.at[i, img_size-7, 5:img_size-5, 0].set(0.9)\n", " \n", " elif label == 3: \n", " images = images.at[i, 5, img_size-10:img_size-5, 0].set(0.9)\n", " images = images.at[i, img_size//2, img_size-10:img_size-5, 0].set(0.9)\n", " images = images.at[i, img_size-6, img_size-10:img_size-5, 0].set(0.9)\n", " images = images.at[i, 5:img_size-5, img_size-6, 0].set(0.9)\n", " \n", " elif label == 4: \n", " images = images.at[i, 5:img_size//2+3, 6, 0].set(0.9)\n", " images = images.at[i, img_size//2, 5:img_size-5, 0].set(0.9)\n", " images = images.at[i, 5:img_size-5, img_size-6, 0].set(0.9)\n", " \n", " elif label == 5: \n", " images = images.at[i, 5, 5:img_size-5, 0].set(0.9)\n", " images = images.at[i, 5:img_size//2+1, 6, 0].set(0.9)\n", " images = images.at[i, img_size//2, 5:img_size-5, 0].set(0.9)\n", " images = images.at[i, img_size//2:img_size-5, img_size-6, 0].set(0.9)\n", " images = images.at[i, img_size-6, 5:img_size-5, 0].set(0.9)\n", " \n", " elif label == 6: \n", " center_x, center_y = img_size//2+3, img_size//2+3\n", " radius = 7\n", " for x in range(img_size):\n", " for y in range(img_size):\n", " dist = (x-center_x)**2 + (y-center_y)**2\n", " if radius**2 - 3 <= dist <= radius**2 + 3 and x >= center_x-2:\n", " images = images.at[i, x, y, 0].set(0.9)\n", " images = images.at[i, 5:img_size-5, 6, 0].set(0.9)\n", " \n", " elif label == 7: \n", " images = images.at[i, 5, 5:img_size-5, 0].set(0.9)\n", " for j in range(15):\n", " images = images.at[i, 6+j, img_size-6-j, 0].set(0.9)\n", " \n", " elif label == 8: \n", " center_x1, center_y1 = img_size//3+1, img_size//2\n", " radius1 = 4\n", " for x in range(img_size):\n", " for y in range(img_size):\n", " dist = (x-center_x1)**2 + (y-center_y1)**2\n", " if radius1**2 - 2 <= dist <= radius1**2 + 2:\n", " images = images.at[i, x, y, 0].set(0.9)\n", " \n", " center_x2, center_y2 = 2*img_size//3-1, img_size//2\n", " radius2 = 4\n", " for x in range(img_size):\n", " for y in range(img_size):\n", " dist = (x-center_x2)**2 + (y-center_y2)**2\n", " if radius2**2 - 2 <= dist <= radius2**2 + 2:\n", " images = images.at[i, x, y, 0].set(0.9)\n", " \n", " elif label == 9:\n", " center_x, center_y = img_size//2-3, img_size//2-3\n", " radius = 7\n", " for x in range(img_size):\n", " for y in range(img_size):\n", " dist = (x-center_x)**2 + (y-center_y)**2\n", " if radius**2 - 3 <= dist <= radius**2 + 3 and x <= center_x+2:\n", " images = images.at[i, x, y, 0].set(0.9)\n", " images = images.at[i, 5:img_size-5, img_size-6, 0].set(0.9) \n", " \n", " # Normalize to [0, 1]\n", " images = (images - images.min()) / (images.max() - images.min())\n", " \n", " return images, labels\n", "\n", "# Generate datasets\n", "print(\"Generating synthetic dataset...\")\n", "X_train, y_train = generate_synthetic_mnist(n_samples=5000)\n", "X_val, y_val = generate_synthetic_mnist(n_samples=500)\n", "X_test, y_test = generate_synthetic_mnist(n_samples=1000)\n", "\n", "print(f\"Training set: {X_train.shape}, {y_train.shape}\")\n", "print(f\"Validation set: {X_val.shape}, {y_val.shape}\")\n", "print(f\"Test set: {X_test.shape}, {y_test.shape}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Visualize Dataset" ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "ExecuteTime": { "end_time": "2025-10-11T10:07:16.196062Z", "start_time": "2025-10-11T10:07:15.864654Z" }, "execution": { "iopub.execute_input": "2026-05-30T17:36:28.685924Z", "iopub.status.busy": "2026-05-30T17:36:28.685720Z", "iopub.status.idle": "2026-05-30T17:36:28.984128Z", "shell.execute_reply": "2026-05-30T17:36:28.983052Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Visualize sample images\n", "fig, axes = plt.subplots(2, 5, figsize=(12, 5))\n", "axes = axes.flatten()\n", "\n", "for i in range(10):\n", " axes[i].imshow(X_train[i, :, :, 0], cmap='gray')\n", " axes[i].set_title(f'Label: {y_train[i]}')\n", " axes[i].axis('off')\n", "\n", "plt.suptitle('Sample Images from Training Set')\n", "plt.tight_layout()\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 2. Build CNN Architecture" ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "ExecuteTime": { "end_time": "2025-10-11T10:20:06.325585Z", "start_time": "2025-10-11T10:20:05.778900Z" }, "execution": { "iopub.execute_input": "2026-05-30T17:36:28.992567Z", "iopub.status.busy": "2026-05-30T17:36:28.992232Z", "iopub.status.idle": "2026-05-30T17:36:33.455947Z", "shell.execute_reply": "2026-05-30T17:36:33.455099Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Input shape: (4, 28, 28, 1)\n", "Output shape: (4, 10)\n", "+-----------------------------+------------+\n", "| Modules | Parameters |\n", "+-----------------------------+------------+\n", "| ('conv1', 'bn', 'weight') | 64 |\n", "| ('conv1', 'conv', 'weight') | 288 |\n", "| ('conv2', 'bn', 'weight') | 128 |\n", "| ('conv2', 'conv', 'weight') | 18.43K |\n", "| ('fc1', 'weight') | 401.54K |\n", "| ('fc2', 'weight') | 1.29K |\n", "| Total | 421.74K |\n", "+-----------------------------+------------+\n" ] }, { "data": { "text/plain": [ "421738" ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" } ], "source": [ "\n", "class ConvBlock(brainstate.nn.Module):\n", " \"\"\"Convolutional block with Conv -> BatchNorm -> ReLU -> MaxPool.\"\"\"\n", "\n", " def __init__(self, in_size, out_channels, kernel_size=3, pool_size=2):\n", " super().__init__()\n", " if isinstance(kernel_size, int):\n", " kernel_size = (kernel_size, kernel_size)\n", " if isinstance(pool_size, int):\n", " pool_size = (pool_size, pool_size)\n", "\n", " self.conv = brainstate.nn.Conv2d(\n", " in_size=in_size,\n", " out_channels=out_channels,\n", " kernel_size=kernel_size,\n", " padding='SAME'\n", " )\n", " self.bn = brainstate.nn.BatchNorm2d(in_size=self.conv.out_size)\n", " self.activation = brainstate.nn.ReLU()\n", " self.pool = brainstate.nn.MaxPool2d(\n", " kernel_size=pool_size,\n", " stride=pool_size,\n", " channel_axis=-1,\n", " in_size=self.conv.out_size\n", " )\n", "\n", " self.in_size = self.conv.in_size\n", " self.out_size = self.pool.out_size\n", "\n", " def update(self, x):\n", " x = self.conv(x)\n", " x = self.bn(x)\n", " x = self.activation(x)\n", " x = self.pool(x)\n", " return x\n", "\n", "\n", "class ImageClassifier(brainstate.nn.Module):\n", " \"\"\"CNN for image classification.\"\"\"\n", "\n", " def __init__(self, input_shape=(28, 28, 1), num_classes=10, dropout_prob=0.5):\n", " super().__init__()\n", " self.conv1 = ConvBlock(in_size=input_shape, out_channels=32, kernel_size=3)\n", " self.conv2 = ConvBlock(in_size=self.conv1.out_size, out_channels=64, kernel_size=3)\n", "\n", " self.flatten = brainstate.nn.Flatten(in_size=self.conv2.out_size)\n", " self.fc1 = brainstate.nn.Linear(in_size=self.flatten.out_size, out_size=(128,))\n", " self.activation = brainstate.nn.ReLU()\n", " self.dropout = brainstate.nn.Dropout(prob=dropout_prob)\n", " self.fc2 = brainstate.nn.Linear(in_size=self.fc1.out_size, out_size=(num_classes,))\n", "\n", " self.in_size = self.conv1.in_size\n", " self.out_size = (num_classes,)\n", "\n", " def update(self, x):\n", " x = self.conv1(x)\n", " x = self.conv2(x)\n", " x = self.flatten(x)\n", " x = self.fc1(x)\n", " x = self.activation(x)\n", " x = self.dropout(x)\n", " x = self.fc2(x)\n", " return x\n", "\n", "# Create model\n", "model = ImageClassifier(num_classes=10)\n", "\n", "# Test forward pass\n", "test_batch = X_train[:4]\n", "with brainstate.environ.context(fit=True):\n", " test_output = model(test_batch)\n", "\n", "print(f\"Input shape: {test_batch.shape}\")\n", "print(f\"Output shape: {test_output.shape}\")\n", "\n", "brainstate.nn.count_parameters(model)\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 3. Training Setup" ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "execution": { "iopub.execute_input": "2026-05-30T17:36:33.458375Z", "iopub.status.busy": "2026-05-30T17:36:33.457896Z", "iopub.status.idle": "2026-05-30T17:36:33.776741Z", "shell.execute_reply": "2026-05-30T17:36:33.775850Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Test loss: 5.3273\n", "Test accuracy: 0.0000\n" ] } ], "source": [ "def cross_entropy_loss(logits, labels):\n", " \"\"\"Compute cross-entropy loss.\n", " \n", " Args:\n", " logits: Model outputs of shape (batch_size, num_classes)\n", " labels: True labels of shape (batch_size,)\n", " \n", " Returns:\n", " Scalar loss value\n", " \"\"\"\n", " # Convert labels to one-hot\n", " num_classes = logits.shape[-1]\n", " one_hot_labels = jax.nn.one_hot(labels, num_classes)\n", " \n", " # Compute log softmax\n", " log_probs = jax.nn.log_softmax(logits, axis=-1)\n", " \n", " # Compute loss\n", " loss = -jnp.mean(jnp.sum(one_hot_labels * log_probs, axis=-1))\n", " \n", " return loss\n", "\n", "def accuracy(logits, labels):\n", " \"\"\"Compute classification accuracy.\n", " \n", " Args:\n", " logits: Model outputs of shape (batch_size, num_classes)\n", " labels: True labels of shape (batch_size,)\n", " \n", " Returns:\n", " Accuracy as a float\n", " \"\"\"\n", " predictions = jnp.argmax(logits, axis=-1)\n", " return jnp.mean(predictions == labels)\n", "\n", "# Test loss and accuracy\n", "test_loss = cross_entropy_loss(test_output, y_train[:4])\n", "test_acc = accuracy(test_output, y_train[:4])\n", "\n", "print(f\"Test loss: {test_loss:.4f}\")\n", "print(f\"Test accuracy: {test_acc:.4f}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Training Step" ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "execution": { "iopub.execute_input": "2026-05-30T17:36:33.778561Z", "iopub.status.busy": "2026-05-30T17:36:33.778388Z", "iopub.status.idle": "2026-05-30T17:36:34.842290Z", "shell.execute_reply": "2026-05-30T17:36:34.841198Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Training step metrics: loss=3.5152, accuracy=0.2812\n" ] } ], "source": [ "\n", "def make_train_step(model, optimizer):\n", " \"\"\"Create a training step function for the model.\"\"\"\n", "\n", " @brainstate.transform.jit\n", " def train_step(x_batch, y_batch):\n", " \"\"\"Perform one training step.\n", "\n", " Args:\n", " x_batch: Input batch\n", " y_batch: Label batch\n", "\n", " Returns:\n", " Dictionary with loss and accuracy\n", " \"\"\"\n", " with brainstate.environ.context(fit=True):\n", " # Define loss function\n", " def loss_fn():\n", " logits = model(x_batch)\n", " return cross_entropy_loss(logits, y_batch)\n", "\n", " # Compute loss and gradients\n", " grads, loss = brainstate.transform.grad(\n", " loss_fn,\n", " model.states(brainstate.ParamState),\n", " return_value=True\n", " )()\n", "\n", " # Update parameters using optimizer\n", " optimizer.update(grads)\n", "\n", " # Compute accuracy\n", " logits = model(x_batch)\n", " acc = accuracy(logits, y_batch)\n", "\n", " return loss, acc\n", "\n", " return train_step\n", "\n", "def make_eval_step(model):\n", " \"\"\"Create an evaluation step function for the model.\"\"\"\n", "\n", " @brainstate.transform.jit\n", " def eval_step(x_batch, y_batch):\n", " \"\"\"Perform one evaluation step.\n", "\n", " Args:\n", " x_batch: Input batch\n", " y_batch: Label batch\n", "\n", " Returns:\n", " Dictionary with loss and accuracy\n", " \"\"\"\n", " with brainstate.environ.context(fit=False):\n", " logits = model(x_batch)\n", " loss = cross_entropy_loss(logits, y_batch)\n", " acc = accuracy(logits, y_batch)\n", "\n", " return loss, acc\n", "\n", " return eval_step\n", "\n", "# Test training step\n", "batch_size = 32\n", "x_batch = X_train[:batch_size]\n", "y_batch = y_train[:batch_size]\n", "\n", "# Create optimizer\n", "optimizer = braintools.optim.Adam(lr=0.001)\n", "optimizer.register_trainable_weights(model.states(brainstate.ParamState))\n", "train_step = make_train_step(model, optimizer)\n", "loss, acc = train_step(x_batch, y_batch)\n", "print(f\"Training step metrics: loss={loss:.4f}, accuracy={acc:.4f}\")\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 4. Complete Training Loop" ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "execution": { "iopub.execute_input": "2026-05-30T17:36:34.844047Z", "iopub.status.busy": "2026-05-30T17:36:34.843811Z", "iopub.status.idle": "2026-05-30T17:37:41.871647Z", "shell.execute_reply": "2026-05-30T17:37:41.870715Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Starting training...\n", "======================================================================\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch 0/20: train_loss=2.0137, train_acc=0.3378, val_loss=2.2757, val_acc=0.2222\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch 5/20: train_loss=0.3559, train_acc=0.8473, val_loss=0.0036, val_acc=1.0000\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch 10/20: train_loss=0.2976, train_acc=0.8627, val_loss=0.0007, val_acc=1.0000\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch 15/20: train_loss=0.2714, train_acc=0.8685, val_loss=0.0001, val_acc=1.0000\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch 19/20: train_loss=0.2525, train_acc=0.8645, val_loss=0.0000, val_acc=1.0000\n", "======================================================================\n", "Training completed in 66.93s\n", "Best validation accuracy: 1.0000 at epoch 3\n" ] } ], "source": [ "def create_batches(X, y, batch_size, shuffle=True):\n", " \"\"\"Create batches from dataset.\n", " \n", " Args:\n", " X: Input data\n", " y: Labels\n", " batch_size: Batch size\n", " shuffle: Whether to shuffle data\n", " \n", " Yields:\n", " Tuples of (x_batch, y_batch)\n", " \"\"\"\n", " n_samples = X.shape[0]\n", " indices = np.arange(n_samples)\n", " \n", " if shuffle:\n", " np.random.shuffle(indices)\n", " \n", " for start_idx in range(0, n_samples, batch_size):\n", " end_idx = min(start_idx + batch_size, n_samples)\n", " batch_indices = indices[start_idx:end_idx]\n", " yield X[batch_indices], y[batch_indices]\n", "\n", "def train_epoch(train_step, X_train, y_train, batch_size):\n", " \"\"\"Train for one epoch.\n", "\n", " Returns:\n", " Dictionary with average metrics\n", " \"\"\"\n", " losses = []\n", " accuracies = []\n", "\n", " for x_batch, y_batch in create_batches(X_train, y_train, batch_size, shuffle=True):\n", " loss, acc = train_step(x_batch, y_batch)\n", " losses.append(float(loss))\n", " accuracies.append(float(acc))\n", "\n", " return {\n", " 'loss': np.mean(losses),\n", " 'accuracy': np.mean(accuracies)\n", " }\n", "\n", "def evaluate(eval_step, X, y, batch_size):\n", " \"\"\"Evaluate model on dataset.\n", "\n", " Returns:\n", " Dictionary with average metrics\n", " \"\"\"\n", " losses = []\n", " accuracies = []\n", "\n", " for x_batch, y_batch in create_batches(X, y, batch_size, shuffle=False):\n", " loss, acc = eval_step(x_batch, y_batch)\n", " losses.append(float(loss))\n", " accuracies.append(float(acc))\n", "\n", " return {\n", " 'loss': np.mean(losses),\n", " 'accuracy': np.mean(accuracies)\n", " }\n", "\n", "# Training configuration\n", "config = {\n", " 'num_epochs': 20,\n", " 'batch_size': 64,\n", " 'learning_rate': 0.001,\n", "}\n", "\n", "# Training history\n", "history = {\n", " 'train_loss': [],\n", " 'train_acc': [],\n", " 'val_loss': [],\n", " 'val_acc': [],\n", "}\n", "\n", "# Recreate model for fresh training\n", "model = ImageClassifier(num_classes=10)\n", "\n", "# Create optimizer\n", "optimizer = braintools.optim.Adam(lr=config['learning_rate'])\n", "optimizer.register_trainable_weights(model.states(brainstate.ParamState))\n", "\n", "# Create train and eval step functions\n", "train_step = make_train_step(model, optimizer)\n", "eval_step = make_eval_step(model)\n", "\n", "print(\"Starting training...\")\n", "print(\"=\" * 70)\n", "\n", "best_val_acc = 0.0\n", "start_time = time.time()\n", "\n", "for epoch in range(config['num_epochs']):\n", " # Train\n", " train_metrics = train_epoch(\n", " train_step, X_train, y_train,\n", " config['batch_size']\n", " )\n", "\n", " # Validate\n", " val_metrics = evaluate(eval_step, X_val, y_val, config['batch_size'])\n", " \n", " # Record history\n", " history['train_loss'].append(train_metrics['loss'])\n", " history['train_acc'].append(train_metrics['accuracy'])\n", " history['val_loss'].append(val_metrics['loss'])\n", " history['val_acc'].append(val_metrics['accuracy'])\n", " \n", " # Track best model\n", " if val_metrics['accuracy'] > best_val_acc:\n", " best_val_acc = val_metrics['accuracy']\n", " best_epoch = epoch\n", " \n", " # Print progress\n", " if epoch % 5 == 0 or epoch == config['num_epochs'] - 1:\n", " print(f\"Epoch {epoch:2d}/{config['num_epochs']}: \"\n", " f\"train_loss={train_metrics['loss']:.4f}, \"\n", " f\"train_acc={train_metrics['accuracy']:.4f}, \"\n", " f\"val_loss={val_metrics['loss']:.4f}, \"\n", " f\"val_acc={val_metrics['accuracy']:.4f}\")\n", "\n", "training_time = time.time() - start_time\n", "print(\"=\" * 70)\n", "print(f\"Training completed in {training_time:.2f}s\")\n", "print(f\"Best validation accuracy: {best_val_acc:.4f} at epoch {best_epoch}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 5. Visualize Training Progress" ] }, { "cell_type": "code", "execution_count": 8, "metadata": { "execution": { "iopub.execute_input": "2026-05-30T17:37:41.873270Z", "iopub.status.busy": "2026-05-30T17:37:41.873094Z", "iopub.status.idle": "2026-05-30T17:37:42.059317Z", "shell.execute_reply": "2026-05-30T17:37:42.058486Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Plot training curves\n", "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))\n", "\n", "# Loss curves\n", "epochs = range(len(history['train_loss']))\n", "ax1.plot(epochs, history['train_loss'], 'b-', label='Train Loss', linewidth=2)\n", "ax1.plot(epochs, history['val_loss'], 'r-', label='Val Loss', linewidth=2)\n", "ax1.set_xlabel('Epoch')\n", "ax1.set_ylabel('Loss')\n", "ax1.set_title('Training and Validation Loss')\n", "ax1.legend()\n", "ax1.grid(True, alpha=0.3)\n", "\n", "# Accuracy curves\n", "ax2.plot(epochs, history['train_acc'], 'b-', label='Train Accuracy', linewidth=2)\n", "ax2.plot(epochs, history['val_acc'], 'r-', label='Val Accuracy', linewidth=2)\n", "ax2.set_xlabel('Epoch')\n", "ax2.set_ylabel('Accuracy')\n", "ax2.set_title('Training and Validation Accuracy')\n", "ax2.legend()\n", "ax2.grid(True, alpha=0.3)\n", "\n", "plt.tight_layout()\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 6. Model Evaluation" ] }, { "cell_type": "code", "execution_count": 9, "metadata": { "execution": { "iopub.execute_input": "2026-05-30T17:37:42.061206Z", "iopub.status.busy": "2026-05-30T17:37:42.061030Z", "iopub.status.idle": "2026-05-30T17:37:42.502353Z", "shell.execute_reply": "2026-05-30T17:37:42.501476Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Test Set Performance:\n", "==================================================\n", "Loss: 0.0003\n", "Accuracy: 1.0000\n", "Error Rate: 0.00%\n" ] } ], "source": [ "# Evaluate on test set\n", "test_metrics = evaluate(eval_step, X_test, y_test, batch_size=64)\n", "\n", "print(\"Test Set Performance:\")\n", "print(\"=\" * 50)\n", "print(f\"Loss: {test_metrics['loss']:.4f}\")\n", "print(f\"Accuracy: {test_metrics['accuracy']:.4f}\")\n", "print(f\"Error Rate: {(1 - test_metrics['accuracy']) * 100:.2f}%\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Confusion Matrix" ] }, { "cell_type": "code", "execution_count": 10, "metadata": { "execution": { "iopub.execute_input": "2026-05-30T17:37:42.504149Z", "iopub.status.busy": "2026-05-30T17:37:42.504001Z", "iopub.status.idle": "2026-05-30T17:37:44.283556Z", "shell.execute_reply": "2026-05-30T17:37:44.282750Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "Per-Class Accuracy:\n", "==================================================\n", "Class 0: 1.0000 (109/109)\n", "Class 1: 1.0000 (90/90)\n", "Class 2: 1.0000 (103/103)\n", "Class 3: 1.0000 (90/90)\n", "Class 4: 1.0000 (100/100)\n", "Class 5: 1.0000 (93/93)\n", "Class 6: 1.0000 (107/107)\n", "Class 7: 1.0000 (98/98)\n", "Class 8: 1.0000 (91/91)\n", "Class 9: 1.0000 (119/119)\n" ] } ], "source": [ "def compute_confusion_matrix(model, X, y, num_classes=10, batch_size=64):\n", " \"\"\"Compute confusion matrix.\n", " \n", " Returns:\n", " Confusion matrix of shape (num_classes, num_classes)\n", " \"\"\"\n", " confusion_matrix = np.zeros((num_classes, num_classes), dtype=int)\n", " \n", " with brainstate.environ.context(fit=False):\n", " for x_batch, y_batch in create_batches(X, y, batch_size, shuffle=False):\n", " logits = model(x_batch)\n", " predictions = jnp.argmax(logits, axis=-1)\n", " \n", " for true_label, pred_label in zip(y_batch, predictions):\n", " confusion_matrix[int(true_label), int(pred_label)] += 1\n", " \n", " return confusion_matrix\n", "\n", "# Compute confusion matrix\n", "cm = compute_confusion_matrix(model, X_test, y_test)\n", "\n", "# Plot confusion matrix\n", "plt.figure(figsize=(10, 8))\n", "plt.imshow(cm, cmap='Blues', interpolation='nearest')\n", "plt.title('Confusion Matrix')\n", "plt.colorbar()\n", "plt.xlabel('Predicted Label')\n", "plt.ylabel('True Label')\n", "\n", "# Add text annotations\n", "for i in range(10):\n", " for j in range(10):\n", " plt.text(j, i, str(cm[i, j]), \n", " ha='center', va='center',\n", " color='white' if cm[i, j] > cm.max() / 2 else 'black')\n", "\n", "plt.tight_layout()\n", "plt.show()\n", "\n", "# Per-class accuracy\n", "print(\"\\nPer-Class Accuracy:\")\n", "print(\"=\" * 50)\n", "for i in range(10):\n", " class_total = cm[i].sum()\n", " class_correct = cm[i, i]\n", " class_acc = class_correct / class_total if class_total > 0 else 0\n", " print(f\"Class {i}: {class_acc:.4f} ({class_correct}/{class_total})\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 7. Visualize Predictions" ] }, { "cell_type": "code", "execution_count": 11, "metadata": { "execution": { "iopub.execute_input": "2026-05-30T17:37:44.286718Z", "iopub.status.busy": "2026-05-30T17:37:44.286533Z", "iopub.status.idle": "2026-05-30T17:37:45.851959Z", "shell.execute_reply": "2026-05-30T17:37:45.851244Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Get predictions for test samples\n", "n_samples = 20\n", "test_samples = X_test[:n_samples]\n", "test_labels = y_test[:n_samples]\n", "\n", "with brainstate.environ.context(fit=False):\n", " logits = model(test_samples)\n", " predictions = jnp.argmax(logits, axis=-1)\n", " probabilities = jax.nn.softmax(logits, axis=-1)\n", "\n", "# Visualize predictions\n", "fig, axes = plt.subplots(4, 5, figsize=(15, 12))\n", "axes = axes.flatten()\n", "\n", "for i in range(n_samples):\n", " ax = axes[i]\n", " \n", " # Show image\n", " ax.imshow(test_samples[i, :, :, 0], cmap='gray')\n", " \n", " # Title with prediction\n", " true_label = int(test_labels[i])\n", " pred_label = int(predictions[i])\n", " confidence = float(probabilities[i, pred_label])\n", " \n", " color = 'green' if true_label == pred_label else 'red'\n", " ax.set_title(\n", " f'True: {true_label}, Pred: {pred_label}\\nConf: {confidence:.2f}',\n", " color=color,\n", " fontsize=10\n", " )\n", " ax.axis('off')\n", "\n", "plt.suptitle('Model Predictions (Green=Correct, Red=Wrong)', fontsize=14)\n", "plt.tight_layout()\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Prediction Confidence Distribution" ] }, { "cell_type": "code", "execution_count": 12, "metadata": { "execution": { "iopub.execute_input": "2026-05-30T17:37:45.854224Z", "iopub.status.busy": "2026-05-30T17:37:45.854043Z", "iopub.status.idle": "2026-05-30T17:37:47.172950Z", "shell.execute_reply": "2026-05-30T17:37:47.171981Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Average confidence for correct predictions: 0.9997\n", "Average confidence for incorrect predictions: nan\n" ] } ], "source": [ "# Analyze prediction confidence\n", "with brainstate.environ.context(fit=False):\n", " all_logits = model(X_test)\n", " all_predictions = jnp.argmax(all_logits, axis=-1)\n", " all_probs = jax.nn.softmax(all_logits, axis=-1)\n", "\n", "# Get max probability for each prediction\n", "max_probs = jnp.max(all_probs, axis=-1)\n", "\n", "# Separate correct and incorrect predictions\n", "correct_mask = (all_predictions == y_test)\n", "correct_probs = max_probs[correct_mask]\n", "incorrect_probs = max_probs[~correct_mask]\n", "\n", "# Plot confidence distributions\n", "plt.figure(figsize=(10, 5))\n", "\n", "plt.hist(correct_probs, bins=20, alpha=0.7, label='Correct Predictions', color='green')\n", "plt.hist(incorrect_probs, bins=20, alpha=0.7, label='Incorrect Predictions', color='red')\n", "\n", "plt.xlabel('Prediction Confidence')\n", "plt.ylabel('Count')\n", "plt.title('Distribution of Prediction Confidence')\n", "plt.legend()\n", "plt.grid(True, alpha=0.3)\n", "\n", "plt.tight_layout()\n", "plt.show()\n", "\n", "print(f\"Average confidence for correct predictions: {jnp.mean(correct_probs):.4f}\")\n", "print(f\"Average confidence for incorrect predictions: {jnp.mean(incorrect_probs):.4f}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 8. Feature Visualization" ] }, { "cell_type": "code", "execution_count": 13, "metadata": { "execution": { "iopub.execute_input": "2026-05-30T17:37:47.175180Z", "iopub.status.busy": "2026-05-30T17:37:47.174958Z", "iopub.status.idle": "2026-05-30T17:37:47.592208Z", "shell.execute_reply": "2026-05-30T17:37:47.591388Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Visualized 16 filters from conv1 (filter shape: 3x3x1)\n" ] } ], "source": [ "# # Visualize learned filters\n", "# def visualize_conv_filters(model, layer_name='conv1'):\n", "# \"\"\"Visualize convolutional filters.\"\"\"\n", "# # Get first conv layer\n", "# conv_layer = getattr(model, layer_name)\n", "\n", "# # Get weight - it might be a dict or direct value\n", "# weight = conv_layer.conv.weight\n", "# if isinstance(weight.value, dict):\n", "# # If it's a dict, get the first value\n", "# filters = list(weight.value.values())[0]\n", "# else:\n", "# filters = weight.value # Shape: (out_ch, in_ch, kh, kw)\n", "\n", "# n_filters = min(16, filters.shape[0])\n", " \n", "# fig, axes = plt.subplots(4, 4, figsize=(10, 10))\n", "# axes = axes.flatten()\n", " \n", "# for i in range(n_filters):\n", "# # Get filter for first input channel\n", "# filter_img = filters[i, 0, :, :]\n", " \n", "# axes[i].imshow(filter_img, cmap='gray')\n", "# axes[i].set_title(f'Filter {i}')\n", "# axes[i].axis('off')\n", " \n", "# plt.suptitle(f'Learned Filters in {layer_name}')\n", "# plt.tight_layout()\n", "# plt.show() \n", "\n", "# visualize_conv_filters(model, 'conv1')\n", "\n", "\n", "\n", "# Visualize learned filters\n", "def visualize_conv_filters(model, layer_name='conv1'):\n", " \"\"\"Visualize convolutional filters.\n", "\n", " Args:\n", " model: The neural network model\n", " layer_name: Name of the convolutional block to visualize (e.g., 'conv1', 'conv2')\n", " \"\"\"\n", " # Get the conv layer\n", " conv_layer = getattr(model, layer_name)\n", "\n", " # Get weight - brainstate stores it in a dict\n", " weight = conv_layer.conv.weight\n", " if isinstance(weight.value, dict):\n", " # Extract the actual weight tensor from dict\n", " filters = list(weight.value.values())[0] \n", " else:\n", " filters = weight.value\n", "\n", " # BrainState Conv2d uses shape: (kernel_height, kernel_width, in_channels, out_channels)\n", " # This is different from PyTorch which uses (out_channels, in_channels, kh, kw)\n", " kh, kw, in_ch, out_ch = filters.shape\n", "\n", " # Visualize up to 16 filters\n", " n_filters = min(16, out_ch)\n", "\n", " fig, axes = plt.subplots(4, 4, figsize=(10, 10))\n", " axes = axes.flatten()\n", "\n", " for i in range(n_filters):\n", " # Get the i-th output filter from the first input channel\n", " # Shape is (kh, kw, in_ch, out_ch), so we need [:, :, 0, i]\n", " filter_img = filters[:, :, 0, i]\n", "\n", " axes[i].imshow(filter_img, cmap='viridis', interpolation='nearest')\n", " axes[i].set_title(f'Filter {i}', fontsize=10)\n", " axes[i].axis('off')\n", "\n", " # Hide any unused subplots\n", " for i in range(n_filters, 16):\n", " axes[i].axis('off')\n", "\n", " plt.suptitle(f'Learned Filters in {layer_name} (shape: {kh}x{kw})', fontsize=14)\n", " plt.tight_layout()\n", " # plt.savefig(f'{layer_name}_filters.png', dpi=150, bbox_inches='tight')\n", " plt.show()\n", "\n", " print(f\"Visualized {n_filters} filters from {layer_name} (filter shape: {kh}x{kw}x{in_ch})\")\n", "\n", "visualize_conv_filters(model, 'conv1')\n", "\n" ] } ], "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": "3.13.11" } }, "nbformat": 4, "nbformat_minor": 4 }