How it works

LLaMA-3.1-8B fine-tuned with LoRA in 4-bit for tool execution under a strict output contract, so downstream code parses the call instead of trusting free text.

Stack

  • PyTorch
  • PEFT / LoRA
  • Hugging Face Transformers
  • 4-bit quantization

Contact / Rourkela, Odisha, India

Write to me