saymrwulf/transformers
1
0
Fork 0
mirror of https://github.com/saymrwulf/transformers.git synced 2026-05-14 20:58:08 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Fix usage of unpad_input function (#35925)

Browse source 
Fix usage of unpad_function

See https://github.com/huggingface/transformers/issues/35899

In the [commit](cdbbe844b1) return type of `unpad_input` was changed.
Now the code support older and newer versions

Co-authored-by: Pavel Gein <pavel.gein@gmail.com>
This commit is contained in:
Pavel Gein 2025-02-06 15:33:42 +05:00 committed by MekkCyber
parent d68779b358
commit 1bfcddbba7
1 changed files with 1 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
src/transformers/modeling_flash_attention_utils.py
View file

@ -121,7 +121,7 @@ def _upad_input(
else:
# The -q_len: slice assumes left padding.
attention_mask = attention_mask[:, -query_length:]
query_layer, indices_q, cu_seqlens_q, max_seqlen_in_batch_q = unpad_input(query_layer, attention_mask)
query_layer, indices_q, cu_seqlens_q, max_seqlen_in_batch_q, *_ = unpad_input(query_layer, attention_mask)
return (
query_layer,