Mehmet Cagri 08e03df940 [Bugfix][ROCm] Use a zero default for masked scales in the MXFP8 GEMM (#59454) ...

Signed-off-by: Mehmet Cagri Kaymak <mehmet.kaymak@amd.com>

2026-10-01 11:32:25 -05:00

..

__init__.py

[Refactor] Move MXFP8 GEMM management into MxFp8LinearKernel (#39205)

2026-04-10 14:02:03 -07:00

b12x.py

[Refactor] Simplify B12X linear kernels and warmup (#52368)

2026-08-17 09:45:45 -07:00

deep_gemm.py

[Model] DeepSeek-V4.1-Flash Model Definitions (#56228)

2026-09-10 20:16:26 +08:00

emulation.py

[Quantization] Let ModelOpt MXFP8 layers load pre-processed weights Purpose (#57316)

2026-09-19 09:16:39 +00:00

flashinfer.py

[Bugfix][Quantization] Fix MXFP8 startup crash on layers below mm_mxfp8 shape limits (#54223)

2026-09-25 02:54:11 +00:00

humming.py

[Feature][Humming] Humming feature integration (#56685)

2026-09-20 08:11:01 -07:00

marlin.py

[Bugfix][Quantization] Fix MXFP8 startup crash on layers below mm_mxfp8 shape limits (#54223)

2026-09-25 02:54:11 +00:00

Mxfp8LinearKernel.py

[Bugfix][Quantization] Fix MXFP8 startup crash on layers below mm_mxfp8 shape limits (#54223)

2026-09-25 02:54:11 +00:00

rocm_block32_gemm.py

[Bugfix][ROCm] Use a zero default for masked scales in the MXFP8 GEMM (#59454)

2026-10-01 11:32:25 -05:00

rocm_native.py

[ROCm] Bump torch 2.13, triton 3.8, torchaudio, torchvision (#50605)

2026-09-25 14:55:52 -05:00

xpu.py

[XPU] Support MXFP8 linear weights for INC DeepSeek V4 model (#48476)

2026-08-06 13:17:48 +08:00