Default Branch

0e3fe896e2 · Support Llama 4 for fused_marlin_moe (#20457) · Updated 2025-07-04 15:55:10 +08:00

Branches

1236aebf0e · Merge remote-tracking branch 'origin/main' into fp8_ep_dp · Updated 2025-06-03 02:53:27 +08:00    vLLM

583
20

5fbbfe9a4c · [BugFix] FA2 MLA Accuracy Issue (#18807) · Updated 2025-05-30 23:50:58 +08:00    vLLM

698
1

2e773e55b3 · docs: merge v1 architecture with class hierarchy · Updated 2025-05-18 14:48:12 +08:00    vLLM

907
1

221118dc85 · [Bugfix] Use a different prompt for benchmark_serving.py test prompt · Updated 2025-05-18 02:36:31 +08:00    vLLM

908
1

f96a3cc713 · test · Updated 2025-05-10 04:31:08 +08:00    vLLM

1092
2

79acf80471 · Fast decode prepare path for prepare_inputs logic · Updated 2025-05-09 01:26:00 +08:00    vLLM

1356
1

bcf3c8230d · Merge branch 'main' into woosuk-jf · Updated 2025-05-05 02:16:07 +08:00    vLLM

1200
3

b73fdb927a · draft · Updated 2025-05-04 01:50:34 +08:00    vLLM

1351
1

3015d5634e · [BugFix][Attention] Fix sliding window attention in V1 giving incorrect results (#17574) · Updated 2025-05-03 02:02:48 +08:00    vLLM

1347
3

c42e8094ec · Skip xgrammar · Updated 2025-04-30 09:27:58 +08:00    vLLM

1304
6

3ed73a0fe5 · Bump actions/setup-python from 5.4.0 to 5.6.0 · Updated 2025-04-28 12:54:55 +08:00    vLLM

1364
1

a7b809e0f0 · Merge remote-tracking branch 'upstream/main' into benchmark-output · Updated 2025-04-23 22:55:50 +08:00    vLLM

1481
8

ec69124eb4 · [Misc] Improve readability of get_open_port function. (#17024) · Updated 2025-04-23 14:16:53 +08:00    vLLM

1489
0
Included

161010c384 · Initial stubs for P/D scheduling changes · Updated 2025-04-19 04:42:49 +08:00    vLLM

1567
1

dc1b4a6f13 · [Core][V0] Enable regex support with xgrammar (#13228) · Updated 2025-04-14 10:13:38 +08:00    vLLM

1656
0
Included

ccd21e1993 · [V1] Fix profiling.py · Updated 2025-04-12 02:36:37 +08:00    vLLM

1685
1

d28ddf8f9f · Address some review comments · Updated 2025-04-08 05:58:43 +08:00    vLLM

1794
8

87e47eb1db · Fix use_ep · Updated 2025-04-08 03:56:41 +08:00    vLLM

1795
1

6de0982dd0 · added · Updated 2025-04-06 22:07:43 +08:00    vLLM

3018
2

296c6572dd · Revert "[V1] DP scale-out (1/N): Use zmq ROUTER/DEALER sockets for input queue (#15906)" · Updated 2025-04-06 12:10:57 +08:00    vLLM

1835
2