Woosuk Kwon 9528e3a05e

[BugFix][Spec Decode] Fix spec token ids in model runner (#20530) ...

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

2025-07-06 19:44:52 +00:00

..

attention

[V1] Support any head size for FlexAttention backend (#20467)

2025-07-06 09:54:36 -07:00

core

Enable V1 for Hybrid SSM/Attention Models (#20016)

2025-07-04 17:46:53 +00:00

engine

[Structured Outputs][V1] Skipping with models doesn't contain tokenizers (#20365)

2025-07-04 15:05:49 +08:00

executor

[Misc] Configurable timeout for execute_model RPC calls via env var (#19544)

2025-06-22 20:36:26 -07:00

metrics

Export NaNs in logits to scheduler_stats if output is corrupted (#18777)

2025-06-20 22:47:16 +08:00

pool

[Misc] Add SPDX-FileCopyrightText (#20428)

2025-07-04 07:40:42 +00:00

sample

[v1] Re-add fp32 support to v1 engine through FlexAttention (#19754)

2025-07-05 09:41:10 +00:00

spec_decode

[V1] Support any head size for FlexAttention backend (#20467)

2025-07-06 09:54:36 -07:00

structured_output

[Structured Outputs][V1] Skipping with models doesn't contain tokenizers (#20365)

2025-07-04 15:05:49 +08:00

worker

[BugFix][Spec Decode] Fix spec token ids in model runner (#20530)

2025-07-06 19:44:52 +00:00

__init__.py

[V1] AsyncLLM Implementation (#9826)

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

kv_cache_interface.py

Enable V1 for Hybrid SSM/Attention Models (#20016)

2025-07-04 17:46:53 +00:00

outputs.py

Export NaNs in logits to scheduler_stats if output is corrupted (#18777)

2025-06-20 22:47:16 +08:00

request.py

[Core] feat: Implement Priority Scheduling in V1 Engine (#19057)

2025-06-22 20:18:08 -07:00

serial_utils.py

[Core] Do not copy array during hashing (#19484)

2025-06-18 15:36:55 -07:00

utils.py

[DP] Support external DP Load Balancer mode (#19790)

2025-07-02 10:21:52 -07:00