423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-16 03:40:34 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix gemma2 attention

Browse Source
This commit is contained in:
hiyouga
2024-07-13 23:33:45 +08:00
parent 7b19e99ed7
commit 2f6af73da2
7 changed files with 53 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/llamafactory/model/model_utils/longlora.py
View File

@@ -326,7 +326,7 @@ def llama_sdpa_attention_forward(
def _apply_llama_patch() -> None:
require_version("transformers>=4.41.2,<=4.42.3", "To fix: pip install transformers>=4.41.2,<=4.42.3")
require_version("transformers>=4.41.2,<=4.42.4", "To fix: pip install transformers>=4.41.2,<=4.42.4")
LlamaAttention.forward = llama_attention_forward
LlamaFlashAttention2.forward = llama_flash_attention_2_forward
LlamaSdpaAttention.forward = llama_sdpa_attention_forward