search

unslothai / unsloth

Finetune Llama 3.1, Mistral, Phi & Gemma LLMs 2-5x faster with 80% less memory
https://unsloth.ai
Apache License 2.0
15.29k stars 1.03k forks source link

support for DeepSeek-Coder-V2-Lite-Base #703

Open Kingatlas115 opened 2 months ago

Kingatlas115 commented 2 months ago

can we lease add suport for DeepSeek-Coder-V2-Lite-Base i want to fine tune this model.

currently tells me its not suported yet.

danielhanchen commented 2 months ago

No sorry currently it's not supported, but with all model support, it will! Sorry about it!

redkit75 commented 3 weeks ago

No sorry currently it's not supported, but with all model support, it will! Sorry about it!

https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Lite-Base/blob/main/modeling_deepseek.py