[cogview4][feat] Support attention mechanism with variable-length support and batch packing #11349
[cogview4] Enhance attention mechanism with variable-length support a…
78b2ad76
[cogview4] Fix rope
1fae35e8
[cogview4] Fix tensor type after qk norm
255cb5af
[cogview4] Add docs for attn processor
1a48dcdd
[chore] Change type hint
f2a6e5ce
Rename as CogView4TrainingAttnProcessor
ccf67523
[refactor] Back to original signature, using `attention_kwargs` instead
fe0c30b5
a-r-r-o-w
approved these changes
on 2025-04-19
Update src/diffusers/models/transformers/transformer_cogview4.py
b70f2080
Apply style fixes
1f657ac6
Merge branch 'main' into main
ddab0087
yiyixuxu
merged
0434db9a
into main 167 days ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub