[SFT] add model_init_kwargs to training_args

huggingface / trl

Train transformer language models with reinforcement learning.

http://hf.co/docs/trl

Apache License 2.0

8.61k stars 1.06k forks source link

[SFT] add model_init_kwargs to training_args #1787

Open kashif opened 2 days ago

kashif commented 2 days ago

Fix SFTTrainer warning by moving model_init_kwargs to the SFTConfig

HuggingFaceDocBuilderDev commented 2 days ago

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.