PyTorch Image Prediction System

This repository contains a simple PyTorch-based image prediction system. It demonstrates how to use a pre-trained AlexNet model from torchvision to classify an image from the web.

Features

• Image Download: Downloads an image from a given URL.
• Image Preprocessing: Applies necessary transformations to prepare the image for the model.
• Model Loading: Uses a pre-trained AlexNet model.
• Inference: Predicts the class of the image and provides the top-5 class probabilities.
• Visualization: Displays the input image and prediction results.

Requirements

• Python 3.6+
• PyTorch 1.6+
• torchvision 0.7+
• numpy
• matplotlib
• PIL

Installation

git clone https://github.com/xPoleStarx/pytorch-image-predict.git
cd pytorch-image-predict

Example Output

The system will output the predicted class for the input image along with the prediction confidence. Additionally, it will display the top-5 class predictions with their respective probabilities.

232: 'Border collie', 23.630033493041992
217: 'English springer, English springer spaniel', 7.991372108459473
176: 'Saluki, gazelle hound', 5.918837547302246
231: 'collie', 5.379868507385254
195: 'Boston bull, Boston terrier', 3.7559683322906494