Dakai An dc0d318177 [Attention] Add FLASH_ATTN_MLA_SPARSE backend for Hopper sparse MLA (#46189) ...

Signed-off-by: Dakai An <dakaian108@gmail.com>

2026-06-24 09:33:10 +00:00

..

attention

[Attention] Add FLASH_ATTN_MLA_SPARSE backend for Hopper sparse MLA (#46189)

2026-06-24 09:33:10 +00:00

core

[Core][DP] Throttle prefills based on local prefill work (#46532)

2026-06-24 02:27:12 +00:00

engine

Stop setting CUDA_VISIBLE_DEVICES internally in vLLM, add device_ids arg (#45026)

2026-06-20 13:38:10 -07:00

executor

Stop setting CUDA_VISIBLE_DEVICES internally in vLLM, add device_ids arg (#45026)

2026-06-20 13:38:10 -07:00

kv_offload

[KV Offloading] Add tiering metric plumbing (#45959)

2026-06-23 15:10:36 +03:00

metrics

[Migration] Migrate GGUF quantization support to plugin (#39612)

2026-06-12 12:02:21 -07:00

pool

[Core] Ensure memory is pinned prior to async h2d copy (#45424)

2026-06-20 20:02:24 -07:00

sample

[Spec Decode] Reject placeholder (-1) draft tokens in rejection sampler (#46533)

2026-06-24 03:32:32 +00:00

simple_kv_offload

[Bugfix][KVConnector] Fix SimpleCPUOffloadConnector GPU->CPU store race (#46278)

2026-06-22 13:08:47 -07:00

spec_decode

[Bugfix][Spec Decode] Fix probabilistic sampling for parallel drafting (#45956)

2026-06-24 05:36:23 +00:00

structured_output

[Core] Ensure memory is pinned prior to async h2d copy (#45424)

2026-06-20 20:02:24 -07:00

worker

[Spec Decode] Reject placeholder (-1) draft tokens in rejection sampler (#46533)

2026-06-24 03:32:32 +00:00

__init__.py

[V1] AsyncLLM Implementation (#9826)

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

cudagraph_dispatcher.py

[Model] Add DiffusionGemma Support (#45163)

2026-06-11 22:17:35 -07:00

kv_cache_interface.py

[Spec Decode] Support mixed KV page sizes for DFlash (#45181)

2026-06-21 22:45:14 +08:00

kv_cache_spec_registry.py

[KVCache] Support Pluggable KVCacheSpec (#37505)

2026-06-03 09:05:16 -07:00

outputs.py

[Core] Cleanup KVConnector handling with PP + fix MRV2 (#43732)

2026-05-28 13:12:03 -07:00

request.py

[Bugfix] Defer block freeing until in-flight steps finish under async scheduling + PD KV consumer (#45357)

2026-06-15 21:36:09 +00:00

serial_utils.py

[Core] Ensure memory is pinned prior to async h2d copy (#45424)

2026-06-20 20:02:24 -07:00

utils.py

[Core] Ensure memory is pinned prior to async h2d copy (#45424)

2026-06-20 20:02:24 -07:00