[FlashInver v0.6.7] Integrate flashinfer_trtllm mxfp8 gemm (#21576)

This commit is contained in:
Ziang Li
2026-04-01 15:55:06 -04:00
committed by GitHub
parent a1c725bdc5
commit a19ef3a615
3 changed files with 75 additions and 28 deletions
@@ -131,5 +131,10 @@ class TestMXFP8GemmFlashinferTrtllm(MXFP8GemmBase, unittest.TestCase):
backend = "flashinfer_trtllm"
@unittest.skipIf(get_device_sm() < 100, "Test requires CUDA SM 100 or higher")
class TestMXFP8GemmFlashinferCutlass(MXFP8GemmBase, unittest.TestCase):
backend = "flashinfer_cutlass"
if __name__ == "__main__":
unittest.main()