README.md

metadata

title: Nimo's Coder Agent
emoji: 💻
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 4.44.0
app_file: app.py
pinned: false
license: mit

Nimo's Personal Coder Agent

A fine-tuned LLM specialized for code generation, debugging, and code review.

Model Details

• Base Model: Qwen2.5-Coder-0.5B-Instruct
• Training Method: QLoRA (4-bit quantization + LoRA)
• Dataset: CodeAlpaca-20k (20,022 examples)
• Training Time: ~5 hours on T4 GPU

Usage

Enter a coding instruction and optionally provide context (existing code for debugging/refactoring).

Links

• GitHub Repository
• Model on HuggingFace

Author

Nimo (CaptainNimo)