[Flashinfer] Support Flashinfer TRTLLM FP8-qkv BF16/FP16-out Attention Kernel (#23647)

Signed-off-by: elvischenv <219235043+elvischenv@users.noreply.github.com>
This commit is contained in:
elvischenv
2025-09-08 20:53:07 -07:00
committed by GitHub
parent b6fbc15634
commit bba1042c6f
5 changed files with 22 additions and 11 deletions
+4 -2
View File
@@ -258,8 +258,10 @@ class AttnFusionPass(VllmInductorPass):
pattern_fp8 = AttentionFp8StaticQuantPattern(layer)
pattern_fp8.register_if_supported(self.patterns)
pattern_nvfp4 = AttentionNvfp4QuantPattern(layer)
pattern_nvfp4.register_if_supported(self.patterns)
if current_platform.is_cuda() and hasattr(torch.ops._C,
"scaled_fp4_quant"):
pattern_nvfp4 = AttentionNvfp4QuantPattern(layer)
pattern_nvfp4.register_if_supported(self.patterns)
if len(attn_layers) == 0:
logger.warning(