4 people 107a03ba63 [Core] Support fp32 lm_head for generation models via head_dtype (RFC #48305 §3.6) (#48390) ...

Signed-off-by: Karthik Kothuri <karthikkothuri2009@gmail.com>
Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: wang.yuqi <yuqi.wang@daocloud.io>

2026-07-13 16:43:34 +08:00

..

attention

[2/N][KV-Cache Layout Refactor] Pack K/V into the content dim across attention backends (#44455)

2026-07-11 11:11:16 -04:00

core

[Bugfix][KV Cache] Don't route uniform-page-size MLA+SWA models into DeepseekV4 packing (#48256)

2026-07-13 08:16:24 +00:00

cudagraph

[MM][Perf][CG] Support dual-path ViT full CUDA graph for DeepSeek-OCR (#43586)

2026-06-16 04:35:20 -07:00

determinism

[CI Test] Mark batch invariance test flaky (#46530)

2026-06-23 21:01:34 +00:00

distributed

[Core] Add prefill step cadence for better non-PD DP balancing (#44558)

2026-06-16 13:17:18 -07:00

e2e

[CI/Build][AMD] Fix ROCm OOM in eagle_correctness_heavy by reserving CUDA graph memory (#47366)

2026-07-10 02:14:38 +00:00

ec_connector

[EC Connector] Add EC Transfer Params (#42433)

2026-07-12 14:35:33 +03:00

engine

Bump Transformers version to 5.10.4 (#41359)

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

executor

[Core][AMD] Propagate shutdown timeout to MultiprocExecutor (#43154)

2026-06-12 15:13:31 -05:00

kv_connector

[2/N][Core] support partial prefix cache hit for hybrid model (#46384)

2026-07-12 05:37:51 +00:00

kv_offload

Make tiering offload region DP-replica aware (#47987)

2026-07-12 13:10:21 +03:00

logits_processors

[Bugfix] Fix thinking_token_budget not enforced after natural </think> re-entry (#45984)

2026-07-10 22:47:51 +00:00

metrics

[CI][Bugfix] Rerun test_engine_log_metrics_ray on Ray GCS startup timeout (#47208)

2026-07-01 21:32:09 -05:00

sample

[Core] Support fp32 lm_head for generation models via head_dtype (RFC #48305 §3.6) (#48390)

2026-07-13 16:43:34 +08:00

shutdown

[ROCm][CI] Make tests/v1/shutdown an importable package (#47085)

2026-06-29 21:01:27 -05:00

simple_kv_offload

[SimpleCPUOffloadConnector] Fix remaining global→block conversions under PCP/DCP (#46394)

2026-06-24 23:05:24 -07:00

spec_decode

Disable dynamic speculative decoding when DP is enabled (#45963)

2026-07-07 12:52:41 +00:00

streaming_input

DCP supports hybrid attention (#40996)

2026-07-09 21:34:45 -07:00

structured_output

[CI] Bump huggingface-hub from v1.10.2 to v1.22.0 (#47551)

2026-07-04 07:45:45 -07:00

tracing

[CI] Fix segfault in tracing test (#47299)

2026-07-01 14:00:37 +00:00

worker

Runtime Draft Weight Update for Speculative Decoding (#46725)

2026-07-11 22:51:53 -07:00

__init__.py

[V1] AsyncLLM Implementation (#9826)

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

test_kv_cache_spec_registry.py

[KVCache] Support Pluggable KVCacheSpec (#37505)

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

test_oracle.py

[Docs] Clean up speculators docs (#34065)

2026-02-18 13:48:11 -08:00

test_outputs.py

[BugFix] Fix spec decoding max_tokens scheduling perf issue (#29542)

2025-11-28 20:52:23 +08:00

test_request.py

[Refactor] Rename WAITING_FOR_FSM to WAITING_FOR_STRUCTURED_OUTPUT_GRAMMAR (#38048)

2026-03-25 11:41:44 -04:00

test_serial_utils.py

Add tensor IPC transfer mechanism for multimodal data (#32104)

2026-03-21 20:10:20 +00:00

test_tensor_ipc_queue.py

[Test] Refactor hard coded device string in test files under compile/quantization/models/model_executor folders (#38901)

2026-04-15 11:02:35 +08:00

utils.py

Convert formatting to use ruff instead of yapf + isort (#26247)

2025-10-05 07:06:22 -07:00