2N2D - Neural Network Development Dashboard

2N2D is an interactive platform to load, visualize, optimize, and export neural network architectures using ONNX and PyTorch, with a web-based dashboard built in Next.js.

---

Features

• Upload and inspect ONNX models
• Load and analyze CSV datasets
• Automatically optimize model architectures
• Export optimized models back to ONNX
• Beautiful, responsive UI powered by Next.js

Project Structure

2n2d/
├── backend/             # FastAPI + PyTorch + ONNX
│   ├── 2n2d.py          # Core logic (model loading, optimization)
│   ├── 2n2denp.py       # FastAPI server with API endpoints
│   └── requirements.txt # Dependencies
├── frontend/            # Next.js frontend interface
│   └── (standard Next.js structure)
└── README.md            # Project documentation

API Overview

Endpoint Method Description

/test - GET - Test server connection /upload-model - POST - Upload ONNX model (binary file) /upload-csv - POST - Upload CSV data (binary file) /optimize - POST - Run optimization (JSON body) /download-optimized - GET Download best model (base64 ONNX)

Project Setup

1. Backend Setup (Python)

cd backend
python -m venv .venv
source .venv/bin/activate  # or .venv\\Scripts\\activate on Windows
pip install -r requirements.txt

2. Frontend Setup (Next.js)

Copy
Edit
cd frontend
npm install
npm run dev