Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.
Python asraudiochineseemotion-recognitionfunasrmcp-servermultilingual-asropenai-compatible-apiparaformerpunctuationpytorchreal-time-asrspeaker-diarizationspeech-recognitionspeech-to-textstreaming-asrtranscriptionvllmvoice-activity-detectionwhisper-alternative
Commits on main
Fix invalid escape sequences in text normalization and models (#3709)
da12182 mehdi B committed 1 month ago
fix: avoid deep-copying the checkpoint in load_pretrained_model (#3729)
9e7f6af YYDH committed 1 month ago
fix(nano): correct punctuation timestamps across VAD merges (#3703)
41778c4 Bruce-Yii committed 1 month ago
docs(nano): clarify training-speed evidence and release scope (#3725)
02f8b43 zhifu gao committed 1 month ago
docs(nano): clarify training speed evidence and release scope
f027d65 zhifu gao committed 1 month ago
fix: use y.device instead of hardcoded .cuda() in data2vec compute_var (#3710)
bad0781 li-lizhe committed 2 months ago
[perf] Fun-ASR-Nano fine-tuning on B200: 0.244 s to 0.068 s per step at 1 GPU, mostly from cuDNN SDPA plan builds on every new batch shape (#3705)
782a7a9 Hexu Zhao committed 2 months ago
fix(paraformer): correct timestamp alpha and peak argument order (#3716)
1878d61 Xu Xiang committed 2 months ago
fix(campplus): bound clustering thread usage and spectrum work (#3699)
b096b79 Gmgge committed 2 months ago
Merge pull request #3714 from modelscope/codex/mixed-device-availability-20260916
615ef05 zhifu gao committed 2 months ago
Merge pull request #3706 from damnwenxi/fix/submodel-device-placement
4d8e087 zhifu gao committed 2 months ago
Merge pull request #3713 from modelscope/codex/growth-target-strict-20260916
6444e20 zhifu gao committed 2 months ago
Merge pull request #3712 from modelscope/codex/roadmap-status-20260916
66eb99a zhifu gao committed 2 months ago
Merge pull request #3711 from modelscope/codex/sensevoice-punctuation-doc-20260915
b07c5fa zhifu gao committed 2 months ago
test(punctuation): guard SenseVoice documentation on exact PR heads
42a7f02 zhifu gao committed 2 months ago
fix: honor submodel device placement across inference calls
aa353d6 damnwenxi committed 2 months ago