[AMD] Use tilelang as default NSA attention backend dispatch on AMD Instinct (#18319)

This commit is contained in:
fxmarty-amd
2026-02-27 01:43:34 -08:00
committed by GitHub
parent 9b2fbf7e6a
commit 9496bbd7b1
2 changed files with 7 additions and 2 deletions
@@ -1477,7 +1477,9 @@ class NativeSparseAttnBackend(
page_size=1,
)
else:
raise ValueError(f"Unsupported {nsa_impl = }")
raise ValueError(
f"Unsupported {nsa_impl = } for forward_extend. Consider using an other attention backend."
)
def forward_decode(
self,