1d_Diffusion_Model_from_scratch.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 600
        },
        "id": "WrYmpHIbdNxa",
        "outputId": "367fa677-fd44-426f-908e-7d9e1f8d3165"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x800 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.optim as optim\n",
        "from torch.utils.data import DataLoader, TensorDataset\n",
        "\n",
        "# Define the mixture of Gaussians\n",
        "def mixture_of_gaussians(num_samples):\n",
        "    means = [-5, 0, 5]\n",
        "    stds = [1, 1, 1]\n",
        "    proportions = [0.3, 0.4, 0.3]\n",
        "    samples = []\n",
        "    for mean, std, prop in zip(means, stds, proportions):\n",
        "        samples.append(np.random.normal(mean, std, int(num_samples * prop)))\n",
        "    return np.concatenate(samples)\n",
        "\n",
        "# Parameters\n",
        "num_samples = 10000\n",
        "timesteps = 1000\n",
        "beta = np.linspace(0.0001, 0.02, timesteps)  # Linear noise schedule\n",
        "\n",
        "# Generate initial data\n",
        "data = mixture_of_gaussians(num_samples)\n",
        "\n",
        "# Forward diffusion process\n",
        "def forward_diffusion(data, beta, timesteps):\n",
        "    noisy_data = []\n",
        "    for t in range(timesteps):\n",
        "        noise = np.random.normal(0, 1, size=data.shape)\n",
        "        data = np.sqrt(1 - beta[t]) * data + np.sqrt(beta[t]) * noise\n",
        "        if t % 100 == 0 or t == timesteps - 1:  # Store every 100th step and the last step\n",
        "            noisy_data.append(data)\n",
        "    return noisy_data\n",
        "\n",
        "# Apply forward diffusion\n",
        "noisy_data = forward_diffusion(data, beta, timesteps)\n",
        "\n",
        "# Visualization of the forward diffusion process\n",
        "def visualize_forward_diffusion(noisy_data):\n",
        "    plt.figure(figsize=(20, 8))\n",
        "    timesteps_to_plot = list(range(0, timesteps, 100)) + [timesteps - 1]\n",
        "    timesteps_to_plot = timesteps_to_plot[:10]  # Ensure we have at most 10 plots\n",
        "    for i, data in enumerate(noisy_data[:10]):\n",
        "        plt.subplot(2, 5, i + 1)  # Adjust the grid to 2 rows and 5 columns\n",
        "        plt.hist(data, bins=100, density=True, alpha=0.6)\n",
        "        plt.title(f'Timestep {timesteps_to_plot[i]}')\n",
        "    plt.tight_layout()\n",
        "    plt.show()\n",
        "\n",
        "# Visualize forward diffusion\n",
        "visualize_forward_diffusion(noisy_data)\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "# Define the neural network for reverse diffusion\n",
        "class DenoiseNN(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(DenoiseNN, self).__init__()\n",
        "        # First fully connected layer: input size is 2 (noisy data + timestep), output size is 128\n",
        "        self.fc1 = nn.Linear(2, 128)\n",
        "        # Second fully connected layer: input size is 128, output size is 128\n",
        "        self.fc2 = nn.Linear(128, 128)\n",
        "        # Third fully connected layer: input size is 128, output size is 1 (predicted noise)\n",
        "        self.fc3 = nn.Linear(128, 1)\n",
        "\n",
        "    def forward(self, x, t):\n",
        "        # Concatenate the noisy data (x) and the normalized timestep (t) along the feature dimension\n",
        "        x = torch.cat((x, t), dim=1)\n",
        "        # Pass the concatenated input through the first layer followed by a ReLU activation function\n",
        "        x = torch.relu(self.fc1(x))\n",
        "        # Pass the output through the second layer followed by a ReLU activation function\n",
        "        x = torch.relu(self.fc2(x))\n",
        "        # Pass the output through the third layer to get the final predicted noise\n",
        "        x = self.fc3(x)\n",
        "        return x\n",
        "\n",
        "# Preparing the dataset for training\n",
        "def create_dataset(data, beta, timesteps):\n",
        "    dataset_x = []  # List to store noisy data at each timestep\n",
        "    dataset_t = []  # List to store timesteps\n",
        "    dataset_y = []  # List to store the true noise added at each timestep\n",
        "\n",
        "    for t in range(timesteps):\n",
        "        # Generate noise for the current timestep\n",
        "        noise = np.random.normal(0, 1, size=data.shape)\n",
        "        # Add noise to the data according to the forward diffusion process\n",
        "        noisy_data = np.sqrt(1 - beta[t]) * data + np.sqrt(beta[t]) * noise\n",
        "        # Append the noisy data to the dataset_x list\n",
        "        dataset_x.append(noisy_data)\n",
        "        # Append the current timestep to the dataset_t list\n",
        "        dataset_t.append(np.full_like(data, t))\n",
        "        # Append the generated noise to the dataset_y list\n",
        "        dataset_y.append(noise)\n",
        "\n",
        "    # Concatenate lists to form arrays for the entire dataset\n",
        "    dataset_x = np.concatenate(dataset_x)\n",
        "    dataset_t = np.concatenate(dataset_t)\n",
        "    dataset_y = np.concatenate(dataset_y)\n",
        "\n",
        "    return dataset_x, dataset_t, dataset_y\n"
      ],
      "metadata": {
        "id": "ki0mWF8idOZr"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "# Create dataset for training\n",
        "dataset_x, dataset_t, dataset_y = create_dataset(data, beta, timesteps)\n",
        "dataset_x = torch.tensor(dataset_x, dtype=torch.float32).view(-1, 1)\n",
        "dataset_t = torch.tensor(dataset_t, dtype=torch.float32).view(-1, 1) / timesteps\n",
        "dataset_y = torch.tensor(dataset_y, dtype=torch.float32).view(-1, 1)\n",
        "\n",
        "train_dataset = TensorDataset(dataset_x, dataset_t, dataset_y)\n",
        "train_loader = DataLoader(train_dataset, batch_size=128, shuffle=True)\n",
        "\n",
        "# Initialize the model, loss function, and optimizer\n",
        "model = DenoiseNN()\n",
        "criterion = nn.MSELoss()\n",
        "optimizer = optim.Adam(model.parameters(), lr=0.001)\n"
      ],
      "metadata": {
        "id": "Ex2xBKFre145"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"x data\",dataset_x[:10])\n",
        "print(\"t data\",dataset_t)\n",
        "print(\"y data\",dataset_y[:10])\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "c5xlK1UCgrzh",
        "outputId": "08e5dd2d-e3c0-4caf-8d15-eecf7adcfe2f"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "x data tensor([[-3.7507],\n",
            "        [-4.2470],\n",
            "        [-5.4270],\n",
            "        [-5.8648],\n",
            "        [-4.8590],\n",
            "        [-4.1878],\n",
            "        [-6.2775],\n",
            "        [-5.3193],\n",
            "        [-5.8124],\n",
            "        [-5.4865]])\n",
            "t data tensor([[0.0000],\n",
            "        [0.0000],\n",
            "        [0.0000],\n",
            "        ...,\n",
            "        [0.9990],\n",
            "        [0.9990],\n",
            "        [0.9990]])\n",
            "y data tensor([[ 1.8801],\n",
            "        [ 0.3958],\n",
            "        [ 1.2670],\n",
            "        [-0.0152],\n",
            "        [-1.1552],\n",
            "        [-0.8710],\n",
            "        [ 1.5027],\n",
            "        [ 0.3172],\n",
            "        [-0.7826],\n",
            "        [ 0.0964]])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "# Training loop\n",
        "num_epochs = 20\n",
        "for epoch in range(num_epochs):\n",
        "    for x, t, y in train_loader:\n",
        "        optimizer.zero_grad()\n",
        "        output = model(x, t)\n",
        "        loss = criterion(output, y)\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "    print(f'Epoch [{epoch+1}/{num_epochs}], Loss: {loss.item():.4f}')\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OO_kKV5Rg0gx",
        "outputId": "8f6385f2-30df-4709-dfdd-597ceb1d0aa1"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch [1/20], Loss: 1.0903\n",
            "Epoch [2/20], Loss: 1.0892\n",
            "Epoch [3/20], Loss: 1.0653\n",
            "Epoch [4/20], Loss: 0.9751\n",
            "Epoch [5/20], Loss: 1.0382\n",
            "Epoch [6/20], Loss: 0.8822\n",
            "Epoch [7/20], Loss: 1.0880\n",
            "Epoch [8/20], Loss: 1.1013\n",
            "Epoch [9/20], Loss: 1.2834\n",
            "Epoch [10/20], Loss: 1.0890\n",
            "Epoch [11/20], Loss: 1.0210\n",
            "Epoch [12/20], Loss: 0.9181\n",
            "Epoch [13/20], Loss: 1.2319\n",
            "Epoch [14/20], Loss: 1.0369\n",
            "Epoch [15/20], Loss: 0.7661\n",
            "Epoch [16/20], Loss: 1.1163\n",
            "Epoch [17/20], Loss: 1.0817\n",
            "Epoch [18/20], Loss: 0.9658\n",
            "Epoch [19/20], Loss: 0.7365\n",
            "Epoch [20/20], Loss: 0.8711\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "# Reverse diffusion process\n",
        "def reverse_diffusion(noise_data, model, beta, timesteps):\n",
        "    data = noise_data[-1]\n",
        "    reverse_data = [data]\n",
        "    for t in reversed(range(timesteps)):\n",
        "        data = torch.tensor(data, dtype=torch.float32).view(-1, 1)\n",
        "        t_tensor = torch.tensor([t/timesteps], dtype=torch.float32).repeat(data.shape[0], 1)\n",
        "        predicted_noise = model(data, t_tensor).detach().numpy().flatten()\n",
        "        data = (data.numpy().flatten() - np.sqrt(beta[t]) * predicted_noise) / np.sqrt(1 - beta[t])\n",
        "        if t % 100 == 0 or t == 0:\n",
        "            reverse_data.append(data)\n",
        "    return reverse_data\n",
        "\n",
        "# Apply reverse diffusion\n",
        "reverse_data = reverse_diffusion(noisy_data, model, beta, timesteps)\n",
        "\n",
        "# Visualization of the reverse diffusion process\n",
        "def visualize_reverse_diffusion(reverse_data):\n",
        "    plt.figure(figsize=(20, 8))\n",
        "    timesteps_to_plot = list(range(timesteps, 0, -100)) + [0]\n",
        "    timesteps_to_plot = timesteps_to_plot[:10]  # Ensure we have at most 10 plots\n",
        "    for i, data in enumerate(reverse_data[:10]):\n",
        "        plt.subplot(2, 5, i + 1)  # Adjust the grid to 2 rows and 5 columns\n",
        "        plt.hist(data, bins=100, density=True, alpha=0.6)\n",
        "        plt.title(f'Timestep {timesteps_to_plot[i]}')\n",
        "    plt.tight_layout()\n",
        "    plt.show()\n",
        "\n",
        "# Visualize reverse diffusion\n",
        "visualize_reverse_diffusion(reverse_data)"
      ],
      "metadata": {
        "id": "UdajbIjqhO-Q"
      },
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "0s0SjVpuRntz"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}