文章摘要

Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU. Language: Python Stars: 5340 Forks: 797