-
ba901536da
fix ep_enabled
main
root
2026-09-03 06:31:50 +00:00
-
c04f54792e
fix ep
root
2026-09-03 06:16:08 +00:00
-
5b646ae665
fix dp arg
root
2026-09-03 05:48:36 +00:00
-
22b60d93a6
trigger build
root
2026-09-03 04:16:24 +00:00
-
10539e25ae
trigger build
root
2026-09-03 03:30:27 +00:00
-
ff8cbeb022
trigger build
root
2026-09-03 02:56:06 +00:00
-
8cf04ebe60
trigger build
root
2026-09-03 02:49:51 +00:00
-
6a12f4aea8
trigger build
root
2026-09-03 02:35:27 +00:00
-
35d63cf0fc
tp = 4, use ep
root
2026-09-03 02:24:32 +00:00
-
d342d8671e
under test, not sure no errors
root
2026-09-03 02:14:32 +00:00
-
041082f5c3
trigger build add dockerfile.ref
root
2026-09-03 01:51:04 +00:00
-
b5dac9c965
DP/EP under test, not sure no errors
root
2026-09-03 01:49:12 +00:00
-
3a50c67884
add dp 2
root
2026-09-02 09:39:03 +00:00
-
84f6b833c0
delete dp 2
root
2026-09-02 09:26:52 +00:00
-
8813233ef2
trigger buiil
root
2026-09-02 09:10:22 +00:00
-
5a38f4a2fa
baseline8 fix d03
root
2026-09-02 09:08:38 +00:00
-
ba380624d0
trigger build
root
2026-09-02 08:55:45 +00:00
-
ae2acbf833
dp 2
root
2026-09-02 08:41:49 +00:00
-
1166c751d8
dp size 2 under test, not sure no errors
root
2026-09-02 08:28:18 +00:00
-
d83ab51908
[fix] </think> to none
root
2026-09-02 06:43:19 +00:00
-
2a85e77c6f
under test, not sure no errors
dylanyunlon
2026-09-02 05:37:05 +00:00
-
9cd06c7a96
under test, not sure no errors
dylanyunlon
2026-09-02 03:01:48 +00:00
-
b54499c607
trigger build
root
2026-09-02 04:38:15 +00:00
-
8a383d6c6b
under test, not sure no errors
DP
2026-09-02 04:36:31 +00:00
-
0c97b7f520
trigger build
root
2026-09-02 02:58:01 +00:00
-
7f16b711a7
fix yaml
root
2026-09-02 02:44:22 +00:00
-
27878c8689
under test, not sure no errors
DP Migration
2026-09-02 01:00:37 +00:00
-
89d522223f
under test, not sure no errors
DP Migration
2026-09-01 15:42:42 +00:00
-
96f4bafcef
under test, not sure no errors
DP Migration
2026-09-01 10:28:11 +00:00
-
94d77cf0b4
under test, not sure no errors
DP Migration
2026-09-01 10:24:14 +00:00
-
8c9d913f3f
add pycache clean
root
2026-09-01 08:53:05 +00:00
-
b168ff9e2d
[perf] xllm_fused_qknorm_rope.so compiled+wired, xllm_cache
root
2026-09-01 06:36:44 +00:00
-
0d88ac4b62
test submit
root
2026-08-25 16:35:45 +00:00
-
628f8ef603
test submit
root
2026-08-25 16:23:46 +00:00
-
2ca835fea7
[fix] timeout error
root
2026-08-25 10:08:10 +00:00
-
f6e2537461
[fix] patch_ops.sh remove envs.py
root
2026-08-25 06:27:50 +00:00
-
1615a313ea
[fix] F.linear→ix_moe_bridge.linear
root
2026-08-25 04:35:36 +00:00
-
b8f54db5e6
[fix] yaml baseline6 format
root
2026-08-25 04:33:29 +00:00
-
3308d9bcf4
[fix] baseline7 9x MoE decode speedup
root
2026-08-25 04:09:43 +00:00
-
1b82bc14a7
[fix] 30s timeout action
root
2026-08-25 04:05:56 +00:00
-
324fd7f3a9
[fix] docker callback timeout
root
2026-08-25 03:54:08 +00:00
-
44314afb70
[fix] remove CUDA compilation, use prebuilt .so only (fixes 30min timeout)
root
2026-08-25 03:29:34 +00:00
-
00654476d3
[fix] baseline7 pyc清除+greedy n>1+fanout宽松+image重试+max_num_seqs env覆盖
root
2026-08-25 02:36:29 +00:00
-
06d56f6ff8
sub791 docker rizhi log
root
2026-08-24 17:01:22 +00:00
-
362da0bef9
[fix] t3: cap default_max_tokens to 8192
root
2026-08-24 09:59:26 +00:00
-
c2edc9a84e
[fix] baseline7 delete null line
root
2026-08-24 07:00:18 +00:00
-
f1fed0d6d5
[fix] baseline7 fix git add qwen3_6_scripts/serving_chat.py!
root
2026-08-24 06:55:36 +00:00
-
ccf22f25e0
[fix] baseline7 sub791 rizhi
root
2026-08-24 06:00:13 +00:00
-
367031b3bc
[fix] t2_n_2 max_num_seqs; t13
root
2026-08-24 04:18:10 +00:00
-
3fe0590266
[fix] baseline5 yaml not found template
root
2026-08-20 14:04:49 +00:00
-
12d2ad55bd
[baseline5] fix ctx_len=2147450879
root
2026-08-20 12:23:50 +00:00
-
56d0c18605
trigger build
root
2026-08-20 10:50:44 +00:00
-
9050014ec3
trigger rebuild
root
2026-08-20 10:02:16 +00:00
-
10151b7c91
[fix] baseline5 _forward_prefix_pytorch 0x7FFF7FFF fix
root
2026-08-20 09:54:28 +00:00
-
fedb48f21e
[feat] baseline5 fused linear+allreduce bridge
root
2026-08-20 08:48:21 +00:00
-
49862ccda4
[fix] baseline5 tribber build
root
2026-08-20 07:37:26 +00:00
-
684d674fc9
[baseline5] fix template
root
2026-08-20 06:48:08 +00:00
-
8842a5ffef
[fix] baseline5 1.8192->4096 solve oom 2. bridge.linear 导过去
root
2026-08-20 06:33:38 +00:00
-
882669b0bc
delete old md
root
2026-08-20 02:28:32 +00:00
-
43509e71b1
[fix] baseline5 delay git add qwen3_6_scripts/patch_ops.sh qwen3_6_scripts/patch_chat_template.py computility-run.yaml! errors
root
2026-08-20 02:26:53 +00:00
-
b7791a0df1
trigger rebuild
root
2026-08-19 16:31:48 +00:00
-
f7352fa03f
[fix] args num
root
2026-08-19 14:01:25 +00:00
-
9a52068533
[fix] baseline5 fused_add_rms_norm 确实是参数数量不匹配。rms_norm 则是参数数量对了但结果没写回去。两个 bug 叠加导致模型输出全损坏
root
2026-08-19 09:57:51 +00:00
-
0de522fb4b
trigger rebuild
root
2026-08-19 08:46:08 +00:00
-
552f1448b9
[md] baseline5 worklog
root
2026-08-19 08:03:09 +00:00
-
79e9208fe4
[fix] baseline5 W2 用 shfl_down 归约
root
2026-08-19 07:25:33 +00:00
-
989bb943ca
trigger rebuild
root
2026-08-19 06:58:35 +00:00
-
858841a1c4
trigger rebuild
root
2026-08-19 06:53:18 +00:00
-
7b0eaf144a
trigger rebuild
root
2026-08-19 06:38:07 +00:00
-
243a79b862
[fix] baseline5 volite warp
root
2026-08-19 06:37:39 +00:00
-
f7e1774cfa
trigger rebuild
root
2026-08-19 05:31:45 +00:00
-
372456bb62
[fix] baseline5 new line delte
root
2026-08-19 05:31:22 +00:00
-
4d63d394ca
[fix] baseline5 fix __shfl_down_sync
root
2026-08-19 05:30:20 +00:00
-
4ec61094b7
[fix] qkv contiguous
root
2026-08-19 04:32:42 +00:00
-
647c018dc1
trigger rebuild
root
2026-08-19 03:39:20 +00:00
-
914dd4a69d
ixf copy_blocks
root
2026-08-19 03:39:03 +00:00
-
bbb347bb0f
trigger rebuild
root
2026-08-19 03:31:13 +00:00
-
f7ed416754
test line
root
2026-08-19 03:24:44 +00:00
-
413e5b1b5f
trigger rebuild
root
2026-08-19 03:10:06 +00:00
-
c8e8473367
test docker build
root
2026-08-19 03:09:44 +00:00
-
7d74ad4e8e
[fix] baseline docker build from ex_engine.python.xllm_ops import _get error
root
2026-08-19 02:46:40 +00:00
-
53daf03d57
trigger rebuild
root
2026-08-19 02:07:52 +00:00
-
cbbc3a6100
[fix] \ No newline at end of file
root
2026-08-19 02:07:29 +00:00
-
8ce55d6e44
trigger rebuild
root
2026-08-19 01:56:55 +00:00
-
679f2c1e29
[fix] baseline4 _xllm_block_copy arg
root
2026-08-19 01:55:54 +00:00
-
5d2faa3c00
[fix] baseline4 ixf to xllm
root
2026-08-19 01:07:16 +00:00
-
c284a27c56
trigger rebuild
root
2026-08-18 17:41:13 +00:00
-
e1870880c5
[fix ]build baseline4
root
2026-08-18 17:40:53 +00:00
-
8196904587
trigger rebuild
root
2026-08-18 17:26:42 +00:00
-
54834c7644
[fix] baseline4 ops_sh fix
root
2026-08-18 17:26:02 +00:00
-
a836ca302a
trigger rebuild
root
2026-08-18 17:05:49 +00:00
-
37a49ee655
[fix] bseline4 sh 添加换行符
root
2026-08-18 17:05:30 +00:00
-
d9b18c5a57
trigger rebuild
root
2026-08-18 15:58:31 +00:00
-
f886510476
trigger rebuild
root
2026-08-18 15:35:33 +00:00
-
06f4ce3bb9
[fix] baseline4 ixf_F.copy_blocks名字不匹配
root
2026-08-18 15:35:03 +00:00
-
eaf4cfa9c1
trigger rebuild
root
2026-08-18 13:50:06 +00:00
-
2579ef6867
[fix] baseline4 yaml
root
2026-08-18 13:43:22 +00:00
-
cb396eca26
trigger rebuild
root
2026-08-18 10:33:20 +00:00
-
8c206feb91
[fix] baseline4 paged_attn.py 的 seq_lens_tensor 包含 0x7FFF7FFF
root
2026-08-18 10:32:55 +00:00
-
8d65615b19
trigger rebuild
root
2026-08-18 10:21:50 +00:00