CoolFace
Apppublic

Hasibur013/cifar10_image_classifier_gradio

sourceHugging Faceupdated 2y agoView on Hugging Face
0likes
App README

🎯 CIFAR-10 Image Classification with Gradio, PyTorch & MLflow

📌 Project Overview

This project builds and deploys an image classification model using the CIFAR-10 dataset. The model is trained using PyTorch Lightning, tracked with MLflow, and deployed as a web app using Gradio.

🚀 Features

  • —Exploratory Data Analysis (EDA) with pandas and seaborn
  • —Custom Data Pipeline using torchvision.transforms
  • —Model Training using PyTorch Lightning
  • —Experiment Tracking with MLflow
  • —Deployment using Gradio
  • —Pretrained Model (ResNet-18) for High Accuracy(85) using simple it was 79
  • —Supports Image Upload in PNG, JPG Formats

🛠️ Tools & Technologies Used

  • —Python 🐍
  • —PyTorch & PyTorch Lightning 🔥
  • —MLflow 🏷️
  • —Gradio 🌐
  • —NumPy & Pandas 📊
  • —Seaborn & Matplotlib 📉

📂 Project Structure

📁 image_classifier
│-- 📂 data
│-- 📂 models
│-- 📂 lightning_logs
│-- 📂 mlruns
│-- app.py   # Gradio Web App
│-- model_build.ipynb  # Model Training Script
│-- requirements.txt
│-- README.md

📸 Model Deployment (Gradio Web App)

Run the following command to launch the web app:

bash
python app.py

The app allows you to upload an image and get predictions from the trained model.

🎥 Project Demo

![Project Demo](https://youtu.be/QcIOw_jK3ag)

🔧 Installation

  1. 1.Clone the repository:
bash
   git clone https://github.com/Hasibur013/cifar10_image_classifier_gradio.git
  1. 1.Navigate to the project folder:
bash
   cd image-classifier
  1. 1.Create virtual Environment:
bash
   Python -m venv venv
  1. 1.Active virtual Environment:
bash
   venv\Scripts\activate
  1. 1.Install dependencies:
bash
   pip install -r requirements.txt

🚀 Train the Model

Run the training script:

bash
notebook model_build.ipynb

📌 Run MLflow Tracking UI

bash
mlflow ui

Then open http://localhost:5000 in your browser to visualize training metrics.

🌟 Future Enhancements

  • —[ ] Support additional image datasets
  • —[ ] Improve model accuracy with hyperparameter tuning
  • —[ ] Deploy as a cloud-based service

🚀 Run Web App

Run the app script:

bash
python app.py

Upload an image (PNG, JPG, etc.) to classify it into one of the CIFAR-10 categories. Image will be airplane, automobile, bird, cat, deer, dog, frog, horse, ship, truck.

Made with ❤️ by Md. Hasibur Rahman