ElizaWszola d9c7730877 [Performance] Extract kv update ops from MLA attention backends (#34627) ...

Signed-off-by: ElizaWszola <ewszola@redhat.com>
Signed-off-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
Co-authored-by: Di Wu <dw2761@nyu.edu>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>

2026-03-02 10:43:19 -05:00

..

__init__.py

[offloader] v2: Hide weight onloading latency via prefetching (#29941)

2026-02-25 17:20:59 -08:00

attention.py

[Attention] FA4 integration (#32974)

2026-03-01 23:44:57 +00:00

cache.py

[offloader] v2: Hide weight onloading latency via prefetching (#29941)

2026-02-25 17:20:59 -08:00

compilation.py

[Performance] Extract kv update ops from MLA attention backends (#34627)

2026-03-02 10:43:19 -05:00

device.py

Turn @config into a dataclass_transform (#31541)

2026-02-03 17:40:59 +00:00

ec_transfer.py

[Misc] Introduce ec_both role EC (encoder cache) connector (#34182)

2026-02-10 18:55:35 +00:00

kernel.py

[UX] Add --moe-backend arg for explicit kernel selection (#33807)

2026-02-25 17:44:44 -08:00

kv_events.py

Turn @config into a dataclass_transform (#31541)

2026-02-03 17:40:59 +00:00

kv_transfer.py

[PD] Change kv_load_failure_policy Default from "recompute" to "fail" (#34896)

2026-02-21 01:34:57 -08:00

load.py

Turn @config into a dataclass_transform (#31541)

2026-02-03 17:40:59 +00:00

lora.py

Turn @config into a dataclass_transform (#31541)

2026-02-03 17:40:59 +00:00

model_arch.py

[Core] Parse vLLM engine required fields from hf_config to model_arch_config (#28454)

2026-01-02 15:13:15 -08:00

model.py

[Misc] Cleanup useless current_platform import (#35715)

2026-03-02 09:36:54 +00:00

multimodal.py

[Refactor] Simplify dummy data generation (#35025)

2026-02-22 20:55:27 -08:00

observability.py

[Feature] OTEL tracing during loading (#31162)

2026-02-05 16:59:28 -08:00

offload.py

[offloader] v2: Hide weight onloading latency via prefetching (#29941)

2026-02-25 17:20:59 -08:00

parallel.py

[EPLB] Enforce sync eplb for NCCL-based all2all backend (#35212)

2026-02-28 05:47:12 +00:00

pooler.py

[Refactor] Pass Renderer to Input Processor (#34329)

2026-02-11 19:38:11 -08:00

profiler.py

Turn @config into a dataclass_transform (#31541)

2026-02-03 17:40:59 +00:00

scheduler.py

[Bug] Refactor max_num_batched_tokens to account for drafting (#34898)

2026-02-22 11:18:46 -05:00

speculative.py

[Model][Spec Decode] Nemotron-H MTP and Mamba Speculative Decoding Support (#33726)

2026-02-24 09:49:56 -08:00

speech_to_text.py

[MM] Allow audio chunking for offline LLM (#34628)

2026-02-23 21:04:28 -08:00

structured_outputs.py

Turn @config into a dataclass_transform (#31541)

2026-02-03 17:40:59 +00:00

utils.py

[Misc] Add deprecated environment variable utilities (#33677)

2026-02-19 21:33:25 -08:00

vllm.py

[torch.compile] Undo the fast_moe_cold_start hack in torch>=2.11 (#35475)

2026-03-01 21:44:22 +00:00

weight_transfer.py

[Feat][RL][2/2] Native Weight Syncing API: IPC (#34171)

2026-02-27 13:45:21 -07:00