Bump flashinfer version to 0.6.14 (#47669)

Signed-off-by: AmeenP <ameenp360@gmail.com>
Signed-off-by: AmeenP <ameen@primeintellect.ai>
Co-authored-by: AmeenP <ameen@primeintellect.ai>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Pavani Majety <pmajety@nvidia.com>
This commit is contained in:
Ameen Patel
2026-07-15 21:00:40 -07:00
committed by GitHub
co-authored by AmeenP Jee Jee Li Claude Pavani Majety
parent df8a0900df
commit ba47bb5be1
6 changed files with 53 additions and 18 deletions
+5
View File
@@ -1075,6 +1075,11 @@ def get_requirements() -> list[str]:
# vllm-flash-attn is built only for CUDA 12.x.
# Skip for other versions.
continue
if "flashinfer-cubin" in req:
# Not on PyPI since 0.6.14 (only https://flashinfer.ai/whl), so
# it cannot be a wheel dependency; flashinfer falls back to
# fetching cubins at runtime when the package is absent.
continue
if "nvidia-cutlass-dsl[cu13]" in req and cuda_major == "12":
# [cu13] extra is the default; strip it on CUDA 12 builds.
req = req.replace("nvidia-cutlass-dsl[cu13]", "nvidia-cutlass-dsl")