mirror of
https://github.com/hanzoai/ml.git
synced 2026-11-19 23:50:51 +00:00
- Created comprehensive Rust training framework for Hanzo ML models - Native integration with zen-agentic-dataset (10.5B tokens) - Support for zen-identity-dataset for personality fine-tuning - CLI tools: hanzo-train and hanzo-eval - Configuration-driven training with YAML/TOML support - LoRA parameter-efficient fine-tuning - Multi-GPU support via CUDA/Metal - Progress tracking and evaluation benchmarks - Template configurations for Qwen, Llama, Mistral models - Complete training pipeline: foundation -> identity -> evaluation - Integration with existing Python zen-trainer framework Features: 🚀 Native Rust performance (no Python overhead) 🎯 Zero-copy data loading for efficient memory usage 🔥 Multi-GPU support via hanzo-ug 📊 Built-in evaluation and benchmarking 🎛️ Flexible YAML/TOML configuration 📈 Progress tracking and metrics Ready for training foundation models on real-world agentic programming data