e5/custon-DNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Problem: Implement a Deep Neural Network\n",
    "\n",
    "### Problem Statement\n",
    "You are tasked with constructing a **Deep Neural Network (DNN)** model to solve a regression task using PyTorch. The objective is to predict target values from synthetic data exhibiting a non-linear relationship.\n",
    "\n",
    "### Requirements\n",
    "Implement the `DNNModel` class that satisfies the following criteria:\n",
    "\n",
    "1. **Model Definition**:\n",
    "   - The model should have:\n",
    "     - An **input layer** connected to a **hidden layer**.\n",
    "     - A **ReLU activation function** for non-linearity.\n",
    "     - An **output layer** with a single unit for regression.\n",
    "\n",
    "<details> <summary>💡 Hint</summary> - Use `nn.Sequential` to simplify the implementation of the `DNNModel`. - Experiment with different numbers of layers and hidden units to optimize performance. - Ensure the final layer has a single output unit (since it's a regression task). </details> <details> <summary>💡 Bonus: Try Custom Loss Functions</summary> Experiment with custom loss functions (e.g., Huber Loss) and compare their performance with MSE. </details>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate synthetic data\n",
    "torch.manual_seed(42)\n",
    "X = torch.rand(100, 2) * 10  # 100 data points with 2 features\n",
    "y = (X[:, 0] + X[:, 1] * 2).unsqueeze(1) + torch.randn(100, 1)  # Non-linear relationship with noise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the Deep Neural Network Model\n",
    "class DNNModel(nn.Module):\n",
    "    def __init__(self):\n",
    "        ...\n",
    "\n",
    "    def forward(self, x):\n",
    "        ...\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [100/1000], Loss: 2.3831\n",
      "Epoch [200/1000], Loss: 0.8492\n",
      "Epoch [300/1000], Loss: 0.7732\n",
      "Epoch [400/1000], Loss: 0.7514\n",
      "Epoch [500/1000], Loss: 0.7371\n",
      "Epoch [600/1000], Loss: 0.7291\n",
      "Epoch [700/1000], Loss: 0.7251\n",
      "Epoch [800/1000], Loss: 0.7233\n",
      "Epoch [900/1000], Loss: 0.7225\n",
      "Epoch [1000/1000], Loss: 0.7221\n",
      "Predictions for [[4.0, 3.0], [7.0, 8.0]]: [[9.915834426879883], [23.081729888916016]]\n"
     ]
    }
   ],
   "source": [
    "# Initialize the model, loss function, and optimizer\n",
    "model = DNNModel()\n",
    "criterion = nn.MSELoss()\n",
    "optimizer = optim.Adam(model.parameters(), lr=0.01)\n",
    "\n",
    "# Training loop\n",
    "epochs = 1000\n",
    "for epoch in range(epochs):\n",
    "    # Forward pass\n",
    "    predictions = model(X)\n",
    "    loss = criterion(predictions, y)\n",
    "\n",
    "    # Backward pass and optimization\n",
    "    optimizer.zero_grad()\n",
    "    loss.backward()\n",
    "    optimizer.step()\n",
    "\n",
    "    # Log progress every 100 epochs\n",
    "    if (epoch + 1) % 100 == 0:\n",
    "        print(f\"Epoch [{epoch + 1}/{epochs}], Loss: {loss.item():.4f}\")\n",
    "\n",
    "# Testing on new data\n",
    "X_test = torch.tensor([[4.0, 3.0], [7.0, 8.0]])\n",
    "with torch.no_grad():\n",
    "    predictions = model(X_test)\n",
    "    print(f\"Predictions for {X_test.tolist()}: {predictions.tolist()}\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.10.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}