InternLM / xtuner

An efficient, flexible and full-featured toolkit for fine-tuning LLM (InternLM2, Llama3, Phi3, Qwen, Mistral, ...)
https://xtuner.readthedocs.io/zh-cn/latest/
Apache License 2.0
3.8k stars 302 forks source link

[Fix] Fix OOM when qlora converting #931

Open fanqiNO1 opened 2 weeks ago

fanqiNO1 commented 2 weeks ago

If the LLM is too big to be loaded in a single GPU, we need device_map = 'auto' to avoid OOM.

According to the issue #715.