ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update llama.py

Browse source
This commit is contained in:
Daniel Han-Chen 2024-02-04 03:08:49 +11:00
commit 36b400e598
1 changed files with 2 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

4
unsloth/models/llama.py
View file

@ -767,7 +767,7 @@ class FastLlamaModel:
@staticmethod
def pre_patch():
LlamaAttention .forward = LlamaAttention_fast_forward
# LlamaAttention .forward = LlamaAttention_fast_forward
LlamaSdpaAttention .forward = LlamaAttention_fast_forward
LlamaFlashAttention2.forward = LlamaAttention_fast_forward
LlamaDecoderLayer .forward = LlamaDecoderLayer_fast_forward
@ -801,7 +801,7 @@ class FastLlamaModel:
f"\ / Bfloat16 = {str(SUPPORTS_BFLOAT16).upper()}. Xformers = {xformers_version}. FA = {HAS_FLASH_ATTENTION}.\n"\
f' "-____-" Free Apache license: http://github.com/unslothai/unsloth'
print(statistics)
# FastLlamaModel.pre_patch()
FastLlamaModel.pre_patch()
if dtype is None:
dtype = torch.float16 if not SUPPORTS_BFLOAT16 else torch.bfloat16