{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "# Automatic Differentiation in PyTorch\n",
        "\n",
        "<!--\n",
        "Project: PyTorch Neural Networks\n",
        "Author: RSK World\n",
        "Website: https://rskworld.in\n",
        "Email: help@rskworld.in\n",
        "Phone: +91 93305 39277\n",
        "-->\n",
        "\n",
        "This notebook demonstrates PyTorch's automatic differentiation feature, which is essential for training neural networks.\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {},
      "outputs": [],
      "source": [
        "# Automatic Differentiation in PyTorch\n",
        "# Project: PyTorch Neural Networks\n",
        "# Author: RSK World\n",
        "# Website: https://rskworld.in\n",
        "# Email: help@rskworld.in\n",
        "# Phone: +91 93305 39277\n",
        "\n",
        "import torch\n",
        "\n",
        "# Enable gradient tracking\n",
        "x = torch.tensor(2.0, requires_grad=True)\n",
        "y = torch.tensor(3.0, requires_grad=True)\n",
        "\n",
        "# Define a function\n",
        "z = x**2 + y**2 + x*y\n",
        "\n",
        "print(f\"x = {x.item()}\")\n",
        "print(f\"y = {y.item()}\")\n",
        "print(f\"z = x² + y² + xy = {z.item()}\")\n",
        "\n",
        "# Compute gradients\n",
        "z.backward()\n",
        "\n",
        "print(f\"\\n∂z/∂x = {x.grad.item()}\")\n",
        "print(f\"∂z/∂y = {y.grad.item()}\")\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## Gradient Descent Example\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {},
      "outputs": [],
      "source": [
        "# Simple gradient descent example\n",
        "# Find the minimum of f(x) = x²\n",
        "\n",
        "x = torch.tensor(5.0, requires_grad=True)\n",
        "learning_rate = 0.1\n",
        "\n",
        "print(\"Gradient Descent to minimize f(x) = x²\")\n",
        "print(f\"Initial x: {x.item()}\")\n",
        "\n",
        "for i in range(10):\n",
        "    # Forward pass\n",
        "    loss = x**2\n",
        "    \n",
        "    # Backward pass\n",
        "    loss.backward()\n",
        "    \n",
        "    # Update x (gradient descent)\n",
        "    with torch.no_grad():\n",
        "        x -= learning_rate * x.grad\n",
        "        x.grad.zero_()\n",
        "    \n",
        "    print(f\"Iteration {i+1}: x = {x.item():.4f}, f(x) = {loss.item():.4f}\")\n",
        "\n",
        "print(f\"\\nFinal x: {x.item():.4f} (should be close to 0)\")\n",
        "\n"
      ]
    }
  ],
  "metadata": {
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 2
}