Thomas Parnell 881e3cbe3b [V1] [Hybrid] Enable piecewise CUDA Graph for mamba layers (#21194) ...

Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>

2025-07-19 19:27:21 +00:00

..

ops

[BugFix] fix 3 issues: (1) using metadata for causal-conv1d, (2) indexing overflow in v1 vLLM, and (3) init_states in v0 (#20838)

2025-07-15 16:08:26 -04:00

__init__.py

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

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

abstract.py

[V1] Enable Mamba2 layers other than MambaMixer2 in the v1 engine (#20660)

2025-07-11 05:53:31 +00:00

mamba2_metadata.py

[Kernel] Triton implementation of causal-conv1d for Mamba-based models (#18218)

2025-07-09 12:53:55 -07:00

mamba_mixer2.py

[V1] [Hybrid] Enable piecewise CUDA Graph for mamba layers (#21194)

2025-07-19 19:27:21 +00:00

mamba_mixer.py

[Kernel] Triton implementation of causal-conv1d for Mamba-based models (#18218)

2025-07-09 12:53:55 -07:00

mamba_utils.py

[V1] [Hybrid] Refactor mamba state shape calculation; enable V1 via cli (#20840)

2025-07-15 04:04:35 -07:00