423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-16 20:00:36 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'hiyouga:main' into main

Browse Source
This commit is contained in:
BUAADreamer
2024-05-15 09:54:14 +08:00
committed by GitHub
parent 60b99f80c2 11bf282dcc
commit 7046144499
20 changed files with 226 additions and 60 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/llmtuner/model/utils/attention.py
View File

@@ -22,7 +22,7 @@ def configure_attn_implementation(config: "PretrainedConfig", model_args: "Model
elif model_args.flash_attn == "sdpa":
if not is_sdpa_available():
logger.warning("Torch>=2.1.1 is required for SDPA attention.")
logger.warning("torch>=2.1.1 is required for SDPA attention.")
return
requested_attn_implementation = "sdpa"
@@ -52,4 +52,4 @@ def print_attn_implementation(config: "PretrainedConfig") -> None:
elif attn_implementation == "sdpa":
logger.info("Using torch SDPA for faster training and inference.")
else:
logger.info("Using vanilla Attention implementation.")
logger.info("Using vanilla attention implementation.")