|
|
a7a2c722af
|
fix large completion probe timeout
|
2026-07-15 22:20:25 +08:00 |
|
|
|
9fc7e98af4
|
fastpath oversized basic test outputs
|
2026-07-15 17:18:39 +08:00 |
|
|
|
2b7880efa7
|
fix platform basic test compatibility
|
2026-07-15 14:25:52 +08:00 |
|
|
|
e374b14bf9
|
add moe bucket profiling and record bottlenecks
|
2026-07-15 11:34:59 +08:00 |
|
|
|
59e497fdc9
|
disable profiling trace by default
|
2026-07-15 04:43:46 +08:00 |
|
|
|
cb962589e9
|
revert moe router topk softmax experiment
|
2026-07-15 04:21:03 +08:00 |
|
|
|
eea0653ace
|
experiment moe router topk softmax
|
2026-07-15 04:19:56 +08:00 |
|
|
|
4b89576cbf
|
reduce docker build context and match working baseline
|
2026-07-15 01:32:21 +08:00 |
|
|
|
af375830c7
|
align platform build config with working submission
|
2026-07-15 00:49:24 +08:00 |
|
|
|
62f10c3375
|
restore official docker build baseline
|
2026-07-14 23:03:24 +08:00 |
|
|
|
2c6cabcc63
|
optimize moe tiny batch tokenwise path
|
2026-07-14 21:49:55 +08:00 |
|
|
|
0597fa6c6d
|
profile decode path and optimize tiny batch moe
|
2026-07-14 20:44:05 +08:00 |
|
|
|
a5abd66e21
|
test eager and custom allreduce toggles
|
2026-07-14 19:18:46 +08:00 |
|
|
|
c8cc9401a4
|
baseline working model service and perf tooling
|
2026-07-14 13:57:40 +08:00 |
|
|
|
1902c81fdd
|
fix issue of loading weight
|
2026-06-30 09:55:13 +08:00 |
|
|
|
f89bc60d59
|
fix multiple issues
|
2026-06-26 17:23:55 +08:00 |
|
|
|
c84151eef9
|
fix issues
|
2026-06-26 12:55:02 +08:00 |
|
|
|
b5806731e0
|
some op overhead optimization
|
2026-06-19 11:19:39 +08:00 |
|
|
|
47a4d9e72a
|
fix no reasoning token issue
|
2026-06-18 12:21:05 +08:00 |
|
|
|
3b8a567e9e
|
fix serving issues when requesting real data
|
2026-06-12 17:57:23 +08:00 |
|
|
|
50e3a05fb0
|
fix incorrect MoE step to ensure decoding speed
|
2026-06-12 11:44:50 +08:00 |
|
|
|
629f878c28
|
initial commit for qwen3.6-moe adaptation
|
2026-06-12 10:10:49 +08:00 |
|
|
|
365da18436
|
Add reasoning parser mechanism + qwen3 parser + bugfixes
|
2026-06-10 18:22:29 +08:00 |
|
|
|
4ab36b51d5
|
Add qwen3_coder tool calling parser
|
2026-06-10 14:38:54 +08:00 |
|
|
|
d972854fb7
|
fix completion token statistic bug when input context is large
|
2026-06-08 15:04:34 +08:00 |
|
|
|
c2de1c83b0
|
Utilize chunked prefill + K-tiling techniques to ensure 100K context
|
2026-06-05 17:00:41 +08:00 |
|
|
|
2d1ef50992
|
chunked prefill support and memory opts
|
2026-06-05 16:03:34 +08:00 |
|
|
|
8c047a70ea
|
some modifications to ensure 50K context input
|
2026-06-04 17:56:29 +08:00 |
|
|
|
1c33ef1355
|
add paged_attn
|
2026-05-29 16:53:39 +08:00 |
|
|
|
3ef8227384
|
initial version of adding chunked attention, ensuring 20K context
|
2026-05-29 16:49:33 +08:00 |
|
|
|
0e89906481
|
Qwen3.6-27B iluvatar bi-v100 adaptation
|
2026-05-21 16:37:24 +08:00 |
|
