Describe the complete pipeline in ML using programming through PyTorch.pdf

•

0 likes•2 views

Describe the complete pipeline in ML using programming through PyTorch. For this, you need to write code that performs linear regression using PyTorch on simulated data. Make sure you include training, testing and evaluation in your code. { "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": null, "metadata": { "id": "V0RhNGmBjFWt" }, "outputs": [], "source": [ "import torch\n", "import numpy as np\n", "import matplotlib.pyplot as plt\n", "import seaborn" ] }, { "cell_type": "code", "source": [ "# Creating a function f(X) with a slope of -5\n", "X = torch.arange(-5, 5, 0.2).view(-1, 1)\n", "func = -5 * X" ], "metadata": { "id": "Vq4OWcCNjJFj" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# Plot the line in red with grids\n", "plt.plot(X.numpy(), func.numpy(), 'r', label='func')\n", "plt.xlabel('x')\n", "plt.ylabel('y')\n", "plt.legend()\n", "plt.grid('True', color='y')\n", "plt.show()" ], "metadata": { "id": "EFDhIKURjPbB" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# Adding Gaussian noise to the function f(X) and saving it in Y\n", "Y = func + 1.7 * torch.randn(X.size())" ], "metadata": { "id": "c2KRcFl5jRuy" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# Plot and visualizing the data points in blue\n", "plt.plot(X.numpy(), Y.numpy(), 'b+', label='Y')\n", "plt.plot(X.numpy(), func.numpy(), 'r', label='func')\n", "plt.xlabel('x')\n", "plt.ylabel('y')\n", "plt.legend()\n", "plt.grid('True', color='y')\n", "plt.show()" ], "metadata": { "id": "ahd47p24jgrB" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# defining the function for forward pass for prediction\n", "def forward(x):\n", " return w * x" ], "metadata": { "id": "T-vtFS9Gjig2" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# evaluating data points with Mean Square Error.\n", "def criterion(y_pred, y):\n", " return torch.mean( (y_pred - y) ** 2 )" ], "metadata": { "id": "DpPaj2Vvjv5A" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "w = torch.tensor(-10.0, requires_grad=True)" ], "metadata": { "id": "UMXfOHC0jx9G" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "step_size = 0.1\n", "loss_list = []\n", "iter = 20" ], "metadata": { "id": "Ml6UCAqujzvm" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "for i in range (iter):\n", " # making predictions with forward pass\n", " Y_pred = forward(X)\n", "\n", "\n", " # calculating the loss between original and predicted data points\n", " loss = criterion(Y_pred, Y)\n", "\n", "\n", " # storing the calculated loss in a list\n", " loss_list.append(loss.item())\n", "\n", "\n", " # backward pass for .

Education

$Describe the complete pipeline in ML using programming through PyTorch. For this, you need to write code that performs linear regression using PyTorch on simulated data. Make sure you include training, testing and evaluation in your code. { "nbformat": 4, "nbformat_minor": 0, "metadata": { "colab": { "provenance": [] }, "kernelspec": { "name": "python3", "display_name": "Python 3" }, "language_info": { "name": "python" } }, "cells": [ { "cell_type": "code", "execution_count": null, "metadata": { "id": "V0RhNGmBjFWt" }, "outputs": [], "source": [ "import torchn", "import numpy as npn", "import matplotlib.pyplot as pltn", "import seaborn" ] }, { "cell_type": "code", "source": [ "# Creating a function f(X) with a slope of -5n", "X = torch.arange(-5, 5, 0.2).view(-1, 1)n", "func = -5 * X" ], "metadata": { "id": "Vq4OWcCNjJFj" }, "execution_count": null, "outputs": []$

$}, { "cell_type": "code", "source": [ "# Plot the line in red with gridsn", "plt.plot(X.numpy(), func.numpy(), 'r', label='func')n", "plt.xlabel('x')n", "plt.ylabel('y')n", "plt.legend()n", "plt.grid('True', color='y')n", "plt.show()" ], "metadata": { "id": "EFDhIKURjPbB" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# Adding Gaussian noise to the function f(X) and saving it in Yn", "Y = func + 1.7 * torch.randn(X.size())" ], "metadata": { "id": "c2KRcFl5jRuy" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# Plot and visualizing the data points in bluen", "plt.plot(X.numpy(), Y.numpy(), 'b+', label='Y')n", "plt.plot(X.numpy(), func.numpy(), 'r', label='func')n", "plt.xlabel('x')n", "plt.ylabel('y')n", "plt.legend()n", "plt.grid('True', color='y')n", "plt.show()" ], "metadata": { "id": "ahd47p24jgrB" }, "execution_count": null,$

$"outputs": [] }, { "cell_type": "code", "source": [ "# defining the function for forward pass for predictionn", "def forward(x):n", " return w * x" ], "metadata": { "id": "T-vtFS9Gjig2" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# evaluating data points with Mean Square Error.n", "def criterion(y_pred, y):n", " return torch.mean( (y_pred - y) ** 2 )" ], "metadata": { "id": "DpPaj2Vvjv5A" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "w = torch.tensor(-10.0, requires_grad=True)" ], "metadata": { "id": "UMXfOHC0jx9G" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "step_size = 0.1n", "loss_list = []n", "iter = 20" ],$

$"metadata": { "id": "Ml6UCAqujzvm" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "for i in range (iter):n", " # making predictions with forward passn", " Y_pred = forward(X)n", "n", "n", " # calculating the loss between original and predicted data pointsn", " loss = criterion(Y_pred, Y)n", "n", "n", " # storing the calculated loss in a listn", " loss_list.append(loss.item())n", "n", "n", " # backward pass for computing the gradients of the loss w.r.t to learnable parametersn", " loss.backward()n", "n", "n", " # updateing the parameters after each iterationn", " w.data = w.data - step_size * w.grad.datan", "n", "n", " # zeroing gradients after each iterationn", " w.grad.data.zero_()n", "n", "n", " # priting the values for understandingn", " print('{},t{},t{}'.format(i, loss.item(), w.item())) n", " n", " n", " # .item() gets the numeric value from the tensor structure" ], "metadata": { "id": "3C11BynFj1yw" }, "execution_count": null, "outputs": [] },$

${ "cell_type": "code", "source": [ "# Plotting the loss after each iterationn", "plt.plot(loss_list, 'r')n", "plt.tight_layout()n", "plt.grid('True', color='y')n", "plt.xlabel("Epochs/Iterations")n", "plt.ylabel("Loss")n", "plt.show()" ], "metadata": { "id": "dt9RKOLSj-EZ" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "w.item()" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "sSswUL6-kQ5O", "outputId": "9412eaed-eaab-4d8c-81fd-822516587349" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "-4.990855693817139" ] }, "metadata": {}, "execution_count": 35 } ] }, { "cell_type": "code", "source": [],$

$"metadata": { "id": "RA8s3FjZkbuU" }, "execution_count": null, "outputs": [] } ] }$