Xin Yang 7a5adad480 [Kernel] Optimize sample_recovered_tokens_kernel (#34974) ...

Signed-off-by: Xin Yang <xyangx@amazon.com>

2026-02-20 19:59:06 -08:00

..

attention

[CI] Revert PRs 34818 and 33600 (#34979)

2026-02-20 13:25:50 -08:00

core

[Scheduler][ASR] Fix CrossAttn blocks per-request for Variable length encoder inputs (#31058)

2026-02-16 11:08:44 +00:00

engine

Support prompt_embeds for pooling requests in output processor (#34904)

2026-02-20 19:57:38 -08:00

executor

[CI] Revert PRs 34818 and 33600 (#34979)

2026-02-20 13:25:50 -08:00

kv_offload

[Bugfix] Fix ARC touch KeyError for non-ready T1 blocks in kv offload (#34576)

2026-02-16 07:33:19 -08:00

metrics

[Renderer] Move InputPreprocessor into Renderer (1/2) (#34510)

2026-02-14 10:14:21 -08:00

pool

[Refactor] Clean up pooler modules (#31897)

2026-01-08 00:07:43 +08:00

sample

[Kernel] Optimize sample_recovered_tokens_kernel (#34974)

2026-02-20 19:59:06 -08:00

spec_decode

[Bugfix] Fix MTP accuracy for GLM-5 (#34385)

2026-02-12 11:08:19 +08:00

structured_output

[Refactor] Simplify BOS/EOS token handling (#34435)

2026-02-12 18:18:24 -08:00

worker

[CI] Revert PRs 34818 and 33600 (#34979)

2026-02-20 13:25:50 -08:00

__init__.py

[V1] AsyncLLM Implementation (#9826)

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

cudagraph_dispatcher.py

[CI] Enable mypy coverage for individual excluded files (#34292)

2026-02-16 07:34:29 -08:00

kv_cache_interface.py

Make mypy opt-out instead of opt-in (#33205)

2026-01-29 09:12:26 +00:00

outputs.py

[ModelRunner V2] Fix spec decoding + logprobs (#33391)

2026-01-31 03:33:26 +00:00

request.py

[Refactor] Simplify BOS/EOS token handling (#34435)

2026-02-12 18:18:24 -08:00

serial_utils.py

[Multimodal] Simplify MM input definitions (#33331)

2026-01-29 13:32:04 +00:00

utils.py

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

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