[MUSA]: Add flashinfer sampling backend (#24978)
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com> Co-authored-by: zhiguo.qin <zhiguo.qin@mthreads.com> Co-authored-by: R0CKSTAR <yeahdongcn@gmail.com>
This commit is contained in:
co-authored by
github-actions[bot]
zhiguo.qin
R0CKSTAR
parent
d9fa84b25b
commit
897587b03a
@@ -108,6 +108,7 @@ sources = [
|
||||
"csrc/memory/weak_ref_tensor.cpp",
|
||||
str(_FLASHINFER_REPO.source_dir / "csrc/norm.cu"),
|
||||
str(_FLASHINFER_REPO.source_dir / "csrc/renorm.cu"),
|
||||
str(_FLASHINFER_REPO.source_dir / "csrc/sampling.cu"),
|
||||
# XXX (MUSA): The following files contain MUSA-specific implementations.
|
||||
"csrc/musa/pos_encoding_contiguous.mu",
|
||||
"csrc/musa/moe_gemv_swiglu.mu",
|
||||
|
||||
Reference in New Issue
Block a user