分支列表

默认分支

main
pre-commit / pre-run-check (push) Has been cancelled
pre-commit / pre-commit (push) Has been cancelled

46e6a83ce1 · [Bugfix][KV Offload] Clean up resources after initialization failure (#51227) · 最后更新于 2026-08-06 11:42:13 +00:00

分支列表

8ad3817a8c · Revert "[ROCm][Perf] Support N=5 in wvSplitK skinny GEMM kernels for speculat…" · 最后更新于 2026-05-29 16:25:29 +00:00    frostbyte_neo

2531
1

0b3ba88f16 · Revert "[CPU] Experimentally enable Triton and MRV2 (#43225)" · 最后更新于 2026-05-29 09:28:43 +00:00    frostbyte_neo

2683
19

7e07848d50 · test: touch llama.py for coverage comparison · 最后更新于 2026-05-28 20:33:50 +00:00    frostbyte_neo

2570
1

d4518933de · test: touch triton_reshape_and_cache_flash.py for coverage comparison · 最后更新于 2026-05-28 20:33:44 +00:00    frostbyte_neo

2570
1

27075fc066 · test: touch jais.py for coverage comparison · 最后更新于 2026-05-28 20:33:42 +00:00    frostbyte_neo

2570
1

f28c233a03 · test: touch tilelang.py for coverage comparison · 最后更新于 2026-05-28 20:33:39 +00:00    frostbyte_neo

2570
1

62409e769a · Update quack-kernels requirement from >=0.3.3 to >=0.4.1 · 最后更新于 2026-05-27 15:28:00 +00:00    frostbyte_neo

2636
1

7607496638 · [CI] Filter import-only files using function-level coverage · 最后更新于 2026-05-26 23:02:20 +00:00    frostbyte_neo

2677
5

bfb22a7d1d · Merge branch 'main' into wentao-fix-flashinfer-layout · 最后更新于 2026-05-26 18:48:26 +00:00    frostbyte_neo

2661
2

7f73e49eda · fix test_spec_decode_acceptance_length for mrv2 · 最后更新于 2026-05-26 15:20:49 +00:00    frostbyte_neo

2671
1

4ac64ec057 · deprecate embed&token_classify · 最后更新于 2026-05-25 13:34:43 +00:00    frostbyte_neo

2691
1

6e2b13a264 · Bump protobuf from 6.33.6 to 7.35.0 · 最后更新于 2026-05-20 03:23:20 +00:00    frostbyte_neo

2838
1

ced1255bad · ci: try remaining 20 gpu_1_queue tests on h200_35gb with NVML fix · 最后更新于 2026-05-19 08:58:57 +00:00    frostbyte_neo

2871
1

c2de222aa3 · Merge branch 'main' into wentao-optimize-model-runner-v2-sampler · 最后更新于 2026-05-18 22:45:59 +00:00    frostbyte_neo

2887
7

154aa68e32 · Fix: keep PYTORCH_NVML_BASED_CUDA_CHECK=1 on MIG · 最后更新于 2026-05-18 08:49:26 +00:00    frostbyte_neo

3052
2

ad7125a431 · [Bugfix] Fix DeepSeek V4 MTP HC state handling (#42320) · 最后更新于 2026-05-15 04:28:34 +00:00    frostbyte_neo

3138
15

42b812edf5 · FIX · 最后更新于 2026-05-13 12:48:02 +00:00    frostbyte_neo

3041
3

a8b0f65f44 · Remove broken allocator pre-warming (initializes CUDA before fork) · 最后更新于 2026-05-13 09:26:43 +00:00    frostbyte_neo

3095
3

ee83fe043f · [CI] Migrate all remaining gpu_1_queue jobs to h200_18gb MIG · 最后更新于 2026-05-12 18:58:58 +00:00    frostbyte_neo

3095
1

593d5a4033 · [Bugfix] Fix mismatched kernel-per-logical blocks in NIXL HMA transfer (#42097) · 最后更新于 2026-05-12 13:53:30 +00:00    frostbyte_neo

3084
0
已包含