Skip to content

FIX: Fix adaption prompt bug with transformers 4.36.0 #1255

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 1 commit into from
Closed

FIX: Fix adaption prompt bug with transformers 4.36.0 #1255

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 6 additions & 1 deletion src/peft/tuners/adaption_prompt/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -73,7 +73,12 @@ def llama_compute_query_states(model: nn.Module, **kwargs) -> torch.Tensor:

seq_len = q_len
if past_key_value is not None:
seq_len += past_key_value[0].shape[-2]
# Newer transformers version (>=4.36.0) have a different cache mechanism
# therefore you need to upack it from the tuple of tuples.
if isinstance(past_key_value[0], tuple):
seq_len += past_key_value[0][0].shape[-2]
else:
seq_len += past_key_value[0].shape[-2]
cos, sin = model.rotary_emb(value_states, seq_len=seq_len)

return llama_apply_rotary_pos_emb(query_states, cos, sin, position_ids)
Expand Down