Harry Mellor e1625498f4

Update where bytes_to_unicode is imported from (#30771) ...

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

2025-12-16 08:05:01 -08:00

..

attention

[ROCm][MTP] Support MTP for AITER MLA backend (#28624)

2025-12-16 14:10:26 +00:00

core

[Bugfix] Whisper fix number of allocated CrossAttn blocks per-request (#30772)

2025-12-16 14:20:19 +00:00

engine

[Bugfix] Fix RequestOutput miss lora_request (#30636)

2025-12-16 01:36:35 -08:00

executor

[Logs] Optimize startup logs 4 (#29903)

2025-12-13 16:12:53 -05:00

kv_offload

CPU KV Offloading: Use more CUDA streams (#29013)

2025-12-14 23:50:45 +00:00

metrics

feat(metrics): Add prefill KV compute metric excluding cached tokens (#30189)

2025-12-09 00:08:48 +00:00

pool

[Model][6/N] Improve all pooling task | Support chunked prefill with ALL pooling (#27145)

2025-12-04 13:44:15 +00:00

sample

[bugfix] fix bug when top_logprobs=0 with spec decoding (#30059)

2025-12-12 09:03:35 -08:00

spec_decode

[ROCm][Bugfix] Add MLACommonMetadata to allowed attention types for speculative decoding (#30430)

2025-12-11 19:25:01 +00:00

structured_output

Update where bytes_to_unicode is imported from (#30771)

2025-12-16 08:05:01 -08:00

worker

[BugFix] Fix memory spike in workspace allocation (#30744)

2025-12-16 06:46:22 -08:00

__init__.py

…

cudagraph_dispatcher.py

[BugFix] Fix assert batch_descriptor.num_tokens == num_tokens_padded (#30173)

2025-12-09 10:36:12 -05:00

kv_cache_interface.py

…

outputs.py

[KVConnector] Add KV events to KV Connectors (#28309)

2025-12-11 15:30:29 +01:00

request.py

[P/D] KV Load Failure Recovery/Abort Configuration (#26813)

2025-12-10 11:00:52 -08:00

serial_utils.py

[Model] Move multimodal_cpu_fields definition to field config (#30181)

2025-12-06 13:40:02 +00:00

utils.py

…