{"data":{"slug":"funaudiollm-sensevoice","name":"SenseVoice","tagline":"Multilingual speech understanding toolkit with ASR, emotion recognition, and audio event detection.","github_url":"https://github.com/FunAudioLLM/SenseVoice","owner":"FunAudioLLM","repo":"SenseVoice","owner_avatar_url":"https://avatars.githubusercontent.com/u/167062371?v=4","primary_language":"C","stars":9332,"forks":827,"topics":["asr","audio-analysis","audio-event-detection","cantonese","cross-lingual","emotion-detection","funasr","language-identification","llama-cpp","multilingual","multilingual-asr","pytorch","sensevoice","speech-emotion-recognition","speech-recognition","speech-to-text","speech-understanding","transcription","voice-ai","whisper-alternative"],"archived":false,"github_pushed_at":"2026-09-10T15:47:44+00:00","maintenance_label":"Active","stars_delta_30d":371,"url":"https://www.graphcanon.com/tools/funaudiollm-sensevoice","markdown_url":"https://www.graphcanon.com/tools/funaudiollm-sensevoice.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/funaudiollm-sensevoice","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=funaudiollm-sensevoice","description":"Open-source SenseVoiceSmall model for Mandarin, Cantonese, English, Japanese, and Korean ASR, language ID, emotion recognition, and audio event detection.","homepage_url":"https://huggingface.co/spaces/FunAudioLLM/SenseVoice","license":"Other","open_issues":7,"watchers":68,"ai_summary":"SenseVoice is a comprehensive multilingual speech understanding software that supports automatic speech recognition (ASR), emotion recognition, and audio event detection in over 50 languages. It offers faster processing speed compared to Whisper while supporting non-autoregressive models.","readme_excerpt":"# Requirements\n\n```shell\npip install -r requirements.txt\n```\n\nSenseVoiceSmall examples and the composed FunASR diarization path require `funasr>=1.3.26`. If you installed this repository earlier, run `pip install -U \"funasr>=1.3.26\"` before retrying the demos.\n\n<a name=\"Usage\"></a>\n\n---\n\n# pip3 install -U funasr funasr-onnx\nfrom pathlib import Path\nfrom funasr_onnx import SenseVoiceSmall\nfrom funasr_onnx.utils.postprocess_utils import rich_transcription_postprocess\n\n\nmodel_dir = \"iic/SenseVoiceSmall\"\n\nmodel = SenseVoiceSmall(model_dir, batch_size=10, quantize=True)\n\n---\n\n### Deployment with FastAPI\n```shell\nexport SENSEVOICE_DEVICE=cuda:0\nfastapi run --port 50000\n```\n\n---\n\n### Requirements\n\n```shell\ngit clone https://github.com/modelscope/FunASR.git && cd FunASR\npip3 install -e ./\n```\n\n---\n\n## 🐳 Docker Support\n\nSenseVoice can be built and run using Docker to simplify setup, ensure reproducibility, and support both CPU and GPU inference.\n\n---\n\n### Build with Docker\n```bash\ndocker build -t sensevoice .\n```\n\n> The build workflow also publishes `ghcr.io/qwenaudio/sensevoice`, but the\n> package is currently private and anonymous pulls return HTTP 401. Use the\n> local build above until the [container package](https://github.com/QwenAudio/SenseVoice/pkgs/container/sensevoice)\n> is marked Public.\n\n---\n\n### Docker Compose\nDocker Compose builds the same image and keeps the model cache on the `sensevoice-models` volume. The default compose file does not request GPUs, so it starts on CPU hosts. For GPU, use the `docker run --gpus all` command above.\n\n---\n\n# CPU is the default. Equivalent: SENSEVOICE_DEVICE=cpu docker compose up --build\n```\n\n---\n\n## License\n\n- Source code in this repository is licensed under the [MIT License](./LICENSE).\n- Model weights are distributed separately and follow the terms on each model card. The official [SenseVoiceSmall model card](https://huggingface.co/FunAudioLLM/SenseVoiceSmall) links to the [FunASR Model Open Source License Agreement](https://github.com/modelscope/FunASR/blob/main/MODEL_LICENSE); other artifacts and conversions may list different terms, so check their model cards before use.\n- The maintainers have provided an [official SenseVoiceSmall license clarification](https://github.com/QwenAudio/SenseVoice/issues/334#issuecomment-5083546605) for the [FunASR Model Open Source License Agreement v1.1](https://github.com/modelscope/FunASR/blob/58830eca4012644aac0c3218c3ccc7d98f003fda/MODEL_LICENSE): Commercial use of the official SenseVoiceSmall weights is permitted when the model license is followed; Section 3 is a responsibility and risk disclaimer rather than an additional non-commercial restriction; and fine-tuned derivative weights may remain private. The Section 2.2 attribution and model-name requirements still apply. This clarification covers the official weights only, so check the terms for third-party conversions and bundled artifacts separately.\n\n<a name=\"Community\"></a>","github_created_at":"2024-07-03T02:56:46+00:00","created_at":"2026-07-11T12:14:47.732433+00:00","updated_at":"2026-09-20T07:16:19.08154+00:00","categories":[{"slug":"speech-audio","name":"Speech & Audio","url":"https://www.graphcanon.com/categories/speech-audio","markdown_url":"https://www.graphcanon.com/categories/speech-audio.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/speech-audio"}],"tags":[{"slug":"asr","name":"asr"},{"slug":"audio-analysis","name":"audio-analysis"},{"slug":"audio-event-detection","name":"audio-event-detection"},{"slug":"cantonese","name":"cantonese"},{"slug":"cross-lingual","name":"cross-lingual"},{"slug":"emotion-detection","name":"emotion-detection"},{"slug":"funasr","name":"funasr"},{"slug":"language-identification","name":"language-identification"}],"trust":{"provenance":{"is_fork":false,"github_id":823429665,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-09-20T07:16:16.836Z","maintenance":{"label":"Active","score":82,"methodology":"github_public_v1","releases_90d":4,"days_since_push":9,"last_release_at":"2026-08-27T13:45:17Z","stars_delta_30d":371,"open_issues_delta_30d":7},"security_summary":{"status":"ok","scanner":"osv@v1","low_count":0,"high_count":0,"last_scan_at":"2026-07-11T12:14:50.601Z","medium_count":0,"scan_profile":"deps","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-09-20T07:16:18.103Z"},"deploy":{"source":"dockerfile:Dockerfile","self_host":true,"observed_at":"2026-09-20T07:16:18.103Z","managed_saas":false},"languages":{"value":["c"],"source":"github.language","observed_at":"2026-09-20T07:16:18.103Z"},"has_docker":{"value":true,"source":"dockerfile:Dockerfile","observed_at":"2026-09-20T07:16:18.103Z"},"license_spdx":{"value":"MIT","source":"github.license","observed_at":"2026-09-20T07:16:18.103Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["Projects requiring speed, with non-autoregressive models being up to 15 times faster than Whisper","Environments where multilingual support is crucial, catering more than 50 languages"],"when_not_to_use":["Applications needing Python-based implementation since SenseVoice relies on C and PyTorch for its framework","Scenarios demanding the lowest level of inference latency as it may have higher overhead compared to simpler models"],"source":"enrich:decision_facts","observed_at":"2026-07-17T01:16:14.757Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"SenseVoice is ideal for applications needing fast multilingual ASR with emotion recognition and audio event detection in over 50 languages."}]}}