Quick Start
wget https://codeanvil.net/codeanvil.shchmod +x codeanvil.sh./codeanvil.sh setup
What is CodeAnvil?
CodeAnvil automates the process of fine-tuning large language models on your specific codebase. The result is an AI assistant that deeply understands your project's architecture, coding patterns, and domain-specific knowledge.
✅ One-Command Setup
Automated installation of all dependencies including PyTorch, LLaMA Factory, and Ollama
🎯 Smart Training
Comprehensive data extraction learns build systems, classes, and architecture
📊 Real-Time Monitoring
Watch training progress with live metrics and beautiful progress bars
🚀 Easy Export
One command exports your model to Ollama for immediate use
Basic Usage
# Train on your codebasecodeanvil train /path/to/myproject# Monitor trainingcodeanvil monitor# Export and usecodeanvil export deepseek-coderollama run codeanvil-deepseek_coder:v1.0
Requirements
- OS: Rocky Linux, Arch Linux, or similar
- GPU: NVIDIA GPU with 16GB+ VRAM (RTX 4080 recommended)
- RAM: 16GB minimum, 32GB+ recommended
- Disk: 100GB free space
- CUDA: 12.1 or later