Skip to content

GitLab

"vllm/vscode:/vscode.git/clone" did not exist on "dde1fa18c9f9ba992a8300a300543d6c18d5f08d"
Unverified Commit cb3935a8 authored by Lucas Wilkinson's avatar Lucas Wilkinson Committed by GitHub
Browse files

[FA4] Update flash-attention to latest upstream FA4 (#38690) 


Signed-off-by: default avatarLucas Wilkinson <lwilkins@redhat.com>
parent 82a006be
Show whitespace changes
Inline Side-by-side
Showing with 3 additions and 3 deletions
cmake/external_projects/vllm_flash_attn.cmake
View file @ cb3935a8
...@@ -39,7 +39,7 @@ else() ...@@ -39,7 +39,7 @@ else()
FetchContent_Declare( FetchContent_Declare(
vllm-flash-attn vllm-flash-attn
GIT_REPOSITORY https://github.com/vllm-project/flash-attention.git GIT_REPOSITORY https://github.com/vllm-project/flash-attention.git
GIT_TAG 29210221863736a08f71a866459e368ad1ac4a95 GIT_TAG c0ec424fd8a546d0cbbf4bf050bbcfe837c55afb
GIT_PROGRESS TRUE GIT_PROGRESS TRUE
# Don't share the vllm-flash-attn build between build types # Don't share the vllm-flash-attn build between build types
BINARY_DIR ${CMAKE_BINARY_DIR}/vllm-flash-attn BINARY_DIR ${CMAKE_BINARY_DIR}/vllm-flash-attn
......
requirements/cuda.txt
View file @ cb3935a8
...@@ -16,5 +16,5 @@ flashinfer-cubin==0.6.7 ...@@ -16,5 +16,5 @@ flashinfer-cubin==0.6.7
nvidia-cudnn-frontend>=1.13.0,<1.19.0 nvidia-cudnn-frontend>=1.13.0,<1.19.0
# QuACK and Cutlass DSL for FA4 (cute-DSL implementation) # QuACK and Cutlass DSL for FA4 (cute-DSL implementation)
nvidia-cutlass-dsl>=4.4.0.dev1 nvidia-cutlass-dsl>=4.4.2
quack-kernels>=0.2.7 quack-kernels>=0.3.3
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment