Conversation
|
/kernel-bot build-and-stage sgl-flash-attn3 |
|
Build request processed. Command: Dispatched (1):
|
|
@adarshxs could you look into the build failures? |
ptxas 13.2 hits an internal compiler error (C7907) when building the
sm90 bf16 aggregators (flash_fwd_hdimall_bf16{,_softcap}_sm90.cu).
CUDA 12.8 and 13.0 build cleanly and the 13.0 build is backwards
compatible, so cap the matrix at 13.0.
|
/kernel-bot build-and-stage sgl-flash-attn3 |
|
Build request processed. Command: Dispatched (1):
|
|
/kernel-bot build-and-stage sgl-flash-attn3 |
|
I can only run |
|
/kernel-bot build-and-stage sgl-flash-attn3 |
|
Build request processed. Command: Dispatched (1):
|
|
@sayakpaul already done in this |
|
/kernel-bot build-and-stage sgl-flash-attn3 |
|
Build request processed. Command: Dispatched (1):
|
|
/kernel-bot build-and-stage sgl-flash-attn3 |
|
Build request processed. Command: Dispatched (1):
|
|
Blocked on #1087 |
|
/kernel-bot build-and-stage sgl-flash-attn3 |
2 similar comments
|
/kernel-bot build-and-stage sgl-flash-attn3 |
|
/kernel-bot build-and-stage sgl-flash-attn3 |
|
Build request processed. Command: Dispatched (1):
|
|
https://huggingface.co/kernels/kernels-staging/sgl-flash-attn3/tree/pr-1054 seems to be working. aarch64 is still failing :/ adarshxs could you test using https://huggingface.co/kernels/kernels-staging/sgl-flash-attn3/tree/pr-1054? I think we will need to provide |
|
@adarshxs a gentle ping |
|
I think the aarch64 builds never worked? https://huggingface.co/kernels/kernels-community/sgl-flash-attn3/tree/v1/build |
What does this PR do?
Ports
sgl-flash-attn3to the Torch stable ABI.Motivation
Changes
Testing
Checklist