🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.
Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU.
Single-GPU LLM fine-tuning preflight: memory estimates, runnable LoRA/QLoRA recipes, and local measurements.
Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)
Multi-LoRA inference server that scales to 1000s of fine-tuned LLMs