Resource-Efficient ML Inference API

A lightweight ML deployment system for compressed image classification.

Features

• MobileNetV2 fine-tuned on CIFAR-10
• ONNX export for deployment
• INT8 quantization for size reduction

Project Structure

MLH_PROJECT/
├── training/          # Training and export scripts
├── models/            # ONNX models
├── app/              # API (coming soon)
└── tests/            # Tests (coming soon)

Setup

pip install -r requirements.txt
cd training
python train.py
python export_onnx.py

MLH Fellowship Application

This project demonstrates production ML deployment for the MLH Fellowship.