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 2024-03-18 03:26:22 +11:00
commit 0c449485b6
1 changed files with 2 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

2
unsloth/models/llama.py
View file

@ -1384,11 +1384,13 @@ class FastLlamaModel:
# Now patch lm_head and embed_tokens
if train_embed_tokens:
print("Unsloth: Casting embed_tokens to float32")
model.model.model.embed_tokens.to(torch.float32)
model.model.model.embed_tokens.requires_grad_(True)
pass
if train_lm_head:
print("Unsloth: Casting lm_head to float32")
model.model.lm_head.to(torch.float32)
model.model.lm_head.requires_grad_(True)
pass