周周周 609f649dd7 [OP] Add flashmla baseline implementation and precision test (#7477)

2026-04-21 13:37:52 +08:00

..

batch_invariant

[Optimization] enable trtllm_all_reduce fusion kernel in glm model (#6660)

2026-04-16 14:10:19 +08:00

benchmarks

…

cache_manager

[Cleanup] Replace torch proxy alias with public compat API (#7348)

2026-04-13 11:43:26 +08:00

ce

…

ci_use

[Others] Fix typo (#7280)

2026-04-14 17:28:22 +08:00

deterministic

…

distributed

[Optimization][DeepSeekV3.2]Reducing slot_mapping compute frequency from twice per layer to a single pre-processing step. (#7367)

2026-04-16 19:54:12 +08:00

e2e

[CI] Add pytest failure log collection and persistence (#7405)

2026-04-16 22:56:17 +08:00

engine

[Optimization] enable trtllm_all_reduce fusion kernel in glm model (#6660)

2026-04-16 14:10:19 +08:00

entrypoints

[Feature] implement log channel separation and request log level system (#7190)

2026-04-16 15:13:05 +08:00

eplb

[Others] Fix typo (#7280)

2026-04-14 17:28:22 +08:00

graph_optimization

[RL] Add clear_graph_opt_backend for glm4_mtp (#7378)

2026-04-15 19:44:15 +08:00

input

[Feature] implement log channel separation and request log level system (#7190)

2026-04-16 15:13:05 +08:00

inter_communicator

[Cleanup] Replace torch proxy alias with public compat API (#7348)

2026-04-13 11:43:26 +08:00

layers

[CI] Remove flashinfer cache cleanup to reduce unit test runtime (#7476)

2026-04-21 11:38:30 +08:00

logger

[Feature] implement log channel separation and request log level system (#7190)

2026-04-16 15:13:05 +08:00

metax_ci

[Others] Fix typo (#7280)

2026-04-14 17:28:22 +08:00

metrics

[BugFix] fix speculative gauge metrics in multi api server (#7082)

2026-03-31 10:52:50 +08:00

model_executor

[Optimization] enable trtllm_all_reduce fusion kernel in glm model (#6660)

2026-04-16 14:10:19 +08:00

model_loader

[Loader] add multi-thread model loading (#6877)

2026-04-09 23:40:15 -07:00

multimodal

[BugFix] fix multimodal hasher hash collision risk when ndarray shape or dtype differs (#7185)

2026-04-08 04:26:02 -07:00

operators

[OP] Add flashmla baseline implementation and precision test (#7477)

2026-04-21 13:37:52 +08:00

output

[Feature] implement log channel separation and request log level system (#7190)

2026-04-16 15:13:05 +08:00

platforms

…

plugins

…

pooling

…

quantization

[Cleanup] Replace torch proxy alias with public compat API (#7348)

2026-04-13 11:43:26 +08:00

reasoning

…

rl

…

router

abort requests (#6992)

2026-03-31 11:02:26 +08:00

scheduler

[Feature] implement log channel separation and request log level system (#7190)

2026-04-16 15:13:05 +08:00

spec_decode

[CI] Reduce execution time for ngram kernel tests (#7242)

2026-04-08 16:54:46 +08:00

splitwise

[Cleanup] Replace torch proxy alias with public compat API (#7348)

2026-04-13 11:43:26 +08:00

trace

…

usage

…

utils

[CI]【Hackathon 10th Spring No.33】config 单测补充 (#6730)

2026-04-09 14:28:54 +08:00

v1

[Cleanup] Replace torch proxy alias with public compat API (#7348)

2026-04-13 11:43:26 +08:00

worker

[Optimization][DeepSeekV3.2]Reducing slot_mapping compute frequency from twice per layer to a single pre-processing step. (#7367)

2026-04-16 19:54:12 +08:00

xpu_ci

[XPU] Split the block_attn operator into smaller operators (#6798)

2026-04-16 14:28:40 +08:00

conftest.py

[CI] Add pytest failure log collection and persistence (#7405)

2026-04-16 22:56:17 +08:00

cov_pytest.ini

[CI] Align with Paddle layer_norm kernel update (#7056)

2026-03-27 22:58:01 +08:00

test_envs.py

…

utils.py

…