Skip to content

[Reland] Attention(23) CUDA #27030

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
titaiwangms merged 5 commits into from
Jan 19, 2026
+745 −396
Merged

[Reland] Attention(23) CUDA #27030

Changes from 1 commit
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
41b0b48
Attention(23) CUDA (#26466)
titaiwangms Jan 15, 2026
cc7a947
set default value to is_output_bnsh
titaiwangms Jan 15, 2026
6384b1d
Merge branch 'main' into titaiwang/reland_attention_cuda_23
titaiwangms Jan 16, 2026
575f79b
Merge branch 'main' into titaiwang/reland_attention_cuda_23
titaiwangms Jan 18, 2026
c84aa34
disable is_causal tests when q_seq != kv_seq
titaiwangms Jan 18, 2026
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
Prev Previous commit
Next Next commit
set default value to is_output_bnsh
  • Loading branch information
@titaiwangms
titaiwangms committed Jan 15, 2026
commit cc7a947285141fe232f356c5cd2088e50218b108
2 changes: 1 addition & 1 deletion onnxruntime/contrib_ops/cpu/bert/attention_parameters.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,7 +34,7 @@ struct AttentionParameters {
float mask_filter_value;
float scale;
bool use_tf32;
bool is_output_bnsh; // whether the output format is BNSH
bool is_output_bnsh = false; // whether the output format is BNSH
AttentionMaskType mask_type;
AttentionQkvFormat qkv_format;
};
Expand Down
Loading