Default Branch

3125d79950 · [Chore] Remove unused PolyNorm layer (#27110) · Updated 2025-10-18 03:03:43 +08:00

Branches

96d5d7b959 · Merge branch 'main' into wentao-optimize-startup-log-2 · Updated 2025-10-18 03:38:43 +08:00

-1
-1

fe4fe2538e · update · Updated 2025-10-18 03:30:32 +08:00

-1
-1

a2599dca0f · fix missing removal · Updated 2025-10-18 02:35:42 +08:00

1498
2

a6427280c1 · [Minor] Remove unnecessary error message · Updated 2025-10-18 02:05:19 +08:00

-1
-1

3565e693c6 · Merge branch 'main' into wentao-fix-mypy-v1 · Updated 2025-10-18 00:53:05 +08:00

-1
-1

69c9a01538 · disable flashinfer warmup · Updated 2025-10-17 00:49:29 +08:00

1551
11

01e389cd94 · fix · Updated 2025-10-17 00:48:51 +08:00

1551
14

6f30ab9ab3 · [Performance] Run shared_experts on a separate cuda stream (in parallel with the FusedMoE) · Updated 2025-10-17 00:10:35 +08:00

-1
-1

c72d44ba4a · Add test for batched triton fallback behavior · Updated 2025-10-16 11:46:02 +08:00

225
7

98e71a4954 · enable all · Updated 2025-10-16 08:01:03 +08:00

1680
12

2797adb329 · cleanup · Updated 2025-10-16 02:07:49 +08:00

1558
2

c3a722fcb2 · [CI Failure] Fix tests with missing TinyLlama-1.1B-Chat-v1.0-FP8-e2e (#26816) · Updated 2025-10-15 02:38:59 +08:00

655
0
Included

38cf8237d4 · Fix pytest verbosity for prime-rl ci · Updated 2025-10-14 09:06:33 +08:00

1648
1

22bf5c5077 · fix · Updated 2025-10-12 02:38:33 +08:00

1694
4

37d0a00b16 · [CI] Skip lm-format-enforcer test cases · Updated 2025-10-11 03:14:25 +08:00

555
1

b8b302cde4 · Update CUDA architecture list in build pipeline for 12.9.1 wheels (#26592) · Updated 2025-10-11 02:15:45 +08:00

2111
34

01efc7ef78 · [ci] fix wheel names for arm wheels (#24898) · Updated 2025-10-08 04:40:13 +08:00

2649
8

944913c0fa · docs: clarify remaining v0 references · Updated 2025-10-07 01:59:13 +08:00

1864
1

920db41128 · [Quantization/NVFP4] Speed up TRTLLM NVFP4 MOE weight loading and fix K/V scale loading for MLA Attn (#25968) · Updated 2025-10-04 04:35:58 +08:00

2380
454

6f62c94d7e · updated · Updated 2025-10-04 01:47:16 +08:00

1934
2