Andreas Karatzas 30fbd05537 [ROCm] Use backend-default dot precision for ReplaySSM (#49909) ...

Signed-off-by: Andreas Karatzas <akaratza@amd.com>

2026-07-27 18:11:06 +08:00

..

configs/selective_state_update

[ROCm] Add tuned selective_state_update config for AMD MI350 (#48159)

2026-07-15 10:18:09 +03:00

cpu

[CPU][Spec Decode] Optimize GDN conv path for speculative decoding (#48577)

2026-07-27 06:20:14 +00:00

gdn_chunk_cutedsl

[M3] Improve indexer for long-context decode (sm100) (#48582)

2026-07-16 18:12:15 -07:00

__init__.py

[Kernel/Model] Migrate mamba_ssm and causal_conv1d kernels to vLLM (#7651)

2024-08-28 15:06:52 -07:00

causal_conv1d.py

[Hardware][CPU] Enable granite-4 model on cpu (#47641)

2026-07-20 06:15:16 +00:00

layernorm_gated.py

replace with torch.cuda.device with with torch.accelerator.device_index (#36144)

2026-03-11 23:12:57 -07:00

mamba_ssm.py

[Hardware][CPU] Enable granite-4 model on cpu (#47641)

2026-07-20 06:15:16 +00:00

replayssm_config.py

[Kernel] ReplaySSM: cache SSM inputs for faster Mamba2 standard decode (#48018)

2026-07-24 09:39:49 -07:00

selective_state_update_replayssm_output_only.py

[ROCm] Use backend-default dot precision for ReplaySSM (#49909)

2026-07-27 18:11:06 +08:00

ssd_bmm.py

[Perf] Add do_not_specialize to Mamba SSD chunk kernels (#43803)

2026-05-28 15:40:02 +00:00

ssd_chunk_scan.py

[Perf] Add do_not_specialize to Mamba SSD chunk kernels (#43803)

2026-05-28 15:40:02 +00:00

ssd_chunk_state.py

[Perf] Add do_not_specialize to Mamba SSD chunk kernels (#43803)

2026-05-28 15:40:02 +00:00

ssd_combined.py

[Hardware][CPU] Enable granite-4 model on cpu (#47641)

2026-07-20 06:15:16 +00:00

ssd_state_passing.py

replace with torch.cuda.device with with torch.accelerator.device_index (#36144)

2026-03-11 23:12:57 -07:00

ssu_dispatch.py

[Hardware][CPU] Enable granite-4 model on cpu (#47641)

2026-07-20 06:15:16 +00:00

triton_helpers.py

[Kernel][Mamba] Optimize Mamba2 SSD prefill Triton kernels (#35397)

2026-03-04 19:47:17 +01:00