423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-10-14 23:58:11 +08:00
Code Issues Packages Projects Releases Wiki Activity

tiny fix

Browse Source 
Former-commit-id: 220d7c1ce15e8013a900e59fe0c7937e38b5c3b5
This commit is contained in:
hiyouga 2024-07-14 10:56:45 +08:00
parent 5ab997d484
commit 71e4404c0d
1 changed files with 2 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
src/llamafactory/model/model_utils/attention.py
View File

@ -36,13 +36,14 @@ def configure_attn_implementation(
if model_args.flash_attn == "auto" or model_args.flash_attn == "fa2":
if is_flash_attn_2_available():
require_version("transformers>=4.42.4", "To fix: pip install transformers>=4.42.4")
require_version("flash_attn>=2.6.0", "To fix: pip install flash_attn>=2.6.0")
logger.warning("Gemma-2 should use flash attention 2, change `flash_attn` to fa2.")
model_args.flash_attn = "fa2"
else:
logger.warning("Gemma-2 should use eager attention, change `flash_attn` to disabled.")
model_args.flash_attn = "disabled"
elif model_args.flash_attn == "sdpa":
raise ValueError("Gemma-2 should use soft-capping attention, while the SDPA attention is not compatible.")
logger.warning("Gemma-2 should use soft-capping attention, while the SDPA attention does not support it.")
if model_args.flash_attn == "auto":
return