p12475398
/
LLaMA-Factory-Mirror
forked from p04798526/LLaMA-Factory-Mirror
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

fix: mixtral output_router_logits

This commit is contained in:
liuzc 2024-04-15 12:11:49 +08:00
parent d1fb6c72b5
commit 9f4fe62386
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
src/llmtuner/model/patcher.py
View File

@ -316,7 +316,7 @@ def patch_config(
if getattr(config, "model_type", None) == "qwen2" and is_trainable and model_args.flash_attn:
setattr(config, "use_cache", False) # qwen2 does not support use_cache when using flashattn
if getattr(config, "model_type", None) == "qwen2_moe" and is_trainable:
if getattr(config, "model_type", None) in ["mixtral", "qwen2_moe"] and is_trainable:
setattr(config, "output_router_logits", True)
init_kwargs["torch_dtype"] = model_args.compute_dtype