Skip to content

Add workaround for default attn_type #578

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

Merged
feifeibear merged 1 commit into from
Oct 21, 2025
Merged

Add workaround for default attn_type #578

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
6 changes: 6 additions & 0 deletions xfuser/core/long_ctx_attention/hybrid/attn_layer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,6 +55,12 @@ def __init__(
attn_type: AttnType = AttnType.FA, the attention type supported inside long context attention, including "TORCH", "FA", "FA3", "SAGE_FP16", "SAGE_FP8"
attn_processor: nn.Module = None, the attention processor can be passed in to replace the attention processor if attn_type is do not support it.
"""

# A workaround to allow running xDiT without having yunchang installed
# while still supporting AttnType.FA as the default value for legacy reasons
if attn_type is None:
attn_type = AttnType.FA

super().__init__(
scatter_idx=scatter_idx,
gather_idx=gather_idx,
Expand Down