Morph MORPH
SIGN IN SIGN UP
dhruvhead / sglang-omni UNCLAIMED
0 0 0 Python

ACTIVITY

Commits Reviews
week month year
COMMITS
20
in the last week
CONTRIBUTORS
14
active
STARS
0
total
FORKS
0
total
TOP CONTRIBUTORS
J
Jiaxin(Jackson) Deng
4 commits
J
Jeffro
2 commits
W
Wenyao Gao
2 commits
L
LG
2 commits
C
Chenye Charlie Zhu
1 commit
赵
赵晨阳
1 commit
M
Makcum888e
1 commit
X
Xinli Jing
1 commit
X
Xinyu Lu
1 commit
问
问问神奇海螺
1 commit
RECENT COMMITS
C
[Qwen3-Omni Perf] Add Breakable Prefill CUDA Graph support for Thinker text/audio prefill (#1381)
Chenye Charlie Zhu
748a0b4
J
[Whisper] Enable long audio chunking for Whisper (#1520)
Jeffro
bcfd457
J
[CI] Register Qwen3-TTS 1.7B in the TTS model rotation (#1505)
Jiaxin(Jackson) Deng
2ed421d
J
[Perf][ASR] Cache the audio front-end constant tables across requests (#1516)
Jiaxin(Jackson) Deng
abc5959
赵
[Model] Support MiniMax Music 3 Model (#1521)
赵晨阳
05e268a
J
fix(transcription): skip no-speech chunks instead of hallucinating text (#1512)
Jiaxin(Jackson) Deng
2260553
M
[NPU] Qwen3-Omni talker support (#1476)
Makcum888e
30b6406
W
[Qwen3-Omni Perf] Default code2wav onto the thinker GPU so the talker gets its own device (#1235)
Wenyao Gao
4ef32ad
X
[Perf] Enable FP8 / CPU offload on TTS pipelines (#1312)
Xinli Jing
19fcd1f
X
Enable MOSS-TD async decode at batch size 1 (#1509)
Xinyu Lu
17face6
L
[Perf] Query MOSS-TD embedding cache before mel extraction (#1495)
LG
302885d
问
[Qwen3-Omni Perf] Vectorize thinker MRoPE; document talker linearization gate (#1149) (#1160)
问问神奇海螺
8d93fdd
J
[ASR Optimization][A-PR3] ARK-ASR: enable shared one-step-lookahead async decode (#1483)
Jiannan L.
00f3bef
D
Whisper ASR W-PR2: prefill coalescing to reduce small-batch prefill (#1485)
Dayuxiaoshui
cc3a9c2
W
[Qwen3-Omni Perf] Graph-safe eager thinker: static hidden capture (#1380)
Wenyao Gao
ee070da
J
[Perf] Qwen3-TTS: cut per-step host work in the AR decode loop (#1462)
Jiaxin(Jackson) Deng
d0873d1
L
perf(moss-td): enable tuned prefill coalescing (#1491)
LG
80805b5
J
perf(qwen3-asr): vectorize decode-step mrope positions (#1472)
Jeffro
fd63f4e
R
[qwen3-asr] add breakable prefill CUDA graph (BCG) support (#1458)
Ratish P
06b1048
J
[Comm] Add rank-to-rank TP KV cache transfers (#1382)
Jingwen Gu
aa6af2f