onnxruntime
Support boolean attention mask in Attention(23) CUDA - MHA case
#27428
Merged
Merged