Jhao-Ting ChenandMichael Goin c2c4c4611a [FIX] fused moe with lora shared expert dual stream (1.07x otps) (#34933) ...

Signed-off-by: Jhao-Ting Chen <jhaotingc@nvidia.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>

2026-02-25 04:40:45 +00:00

..

layers

[FIX] fused moe with lora shared expert dual stream (1.07x otps) (#34933)

2026-02-25 04:40:45 +00:00

ops

[LoRA] Update LoRA expand kernel block_n calculation (#32621)

2026-02-23 23:17:53 -08:00

punica_wrapper

[Bug Fix] Fix naive_block_assignment always defaulting to False due to arg misalignment (#33848)

2026-02-12 11:30:57 +08:00

__init__.py

[Experimental] Add multi-LoRA support (#1804)

2024-01-23 15:26:37 -08:00

lora_model.py

[LoRA][Spec Decode] Support LoRA for Nemotron-H MTP models (#32265)

2026-01-27 07:53:26 -08:00

lora_weights.py

[Misc] Replace Optional[X] with X | None syntax (#33332)

2026-01-30 01:56:59 -08:00

model_manager.py

[Tiny] Rename encoder budget file to more specific name (#34103)

2026-02-09 03:48:19 +00:00

peft_helper.py

Update Optional[x] -> x | None and Union[x, y] to x | y (#26633)

2025-10-12 09:51:31 -07:00

request.py

[Feat] allow inplace loading lora (#31326)

2026-01-20 10:15:20 +08:00

resolver.py

Update Optional[x] -> x | None and Union[x, y] to x | y (#26633)

2025-10-12 09:51:31 -07:00

utils.py

Reduce the kernel overhead when num of active loras is smaller than max loras. Multiple cuda graphs are captured for each num of active-loras. (#32005)

2026-02-02 12:30:06 -05:00

worker_manager.py

[LoRA][Spec Decode] Support LoRA for Nemotron-H MTP models (#32265)

2026-01-27 07:53:26 -08:00