transformers
a65bf6c9
- Support FA2 flash_attn_with_kvcache for XPU continuous batching (#46028)
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
110 days ago
Support FA2 flash_attn_with_kvcache for XPU continuous batching (#46028) * Support FA2 flash_attn_with_kvcache for XPU continuous batching * Update according to the comments * simplify the code * Code quality
References
#46028 - Support FA2 flash_attn_with_kvcache for XPU continuous batching
Author
YangKai0616
Parents
70257e9a
Loading