Woosuk Kwon 98f47f2a40 [V1] Optimize the CPU overheads in FlashAttention custom op (#10733) ...

Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>

2024-11-28 09:01:02 -08:00

..

attention

[V1] Optimize the CPU overheads in FlashAttention custom op (#10733)

2024-11-28 09:01:02 -08:00

core

[V1] Do not allocate beyond the max_model_len (#10730)

2024-11-28 00:13:15 -08:00

engine

[V1] Enable profile for LLMEngine (#10665)

2024-11-26 10:36:45 +00:00

executor

[V1] Fix Configs (#9971)

2024-11-04 00:24:40 +00:00

sample

[V1] Support per-request seed (#9945)

2024-11-03 09:14:17 -08:00

worker

[V1] Refactor model executable interface for multimodal models (#10570)

2024-11-26 20:46:11 +00:00

__init__.py

[V1] AsyncLLM Implementation (#9826)

2024-11-11 23:05:38 +00:00

outputs.py

[V1] Implement vLLM V1 [1/N] (#9289)

2024-10-22 01:24:07 -07:00

request.py

[1/N] Initial prototype for multi-modal processor (#10044)

2024-11-13 12:39:03 +00:00

serial_utils.py

[V1] Use pickle for serializing EngineCoreRequest & Add multimodal inputs to EngineCoreRequest (#10245)

2024-11-12 08:57:14 -08:00

utils.py

[V1] Add all_token_ids attribute to Request (#10135)

2024-11-07 17:08:24 -08:00