answered from live GitHub & Hugging Face data
I want to find model to analyze music audio, figure out measure, bpm, etc
🤗 laion/clap-htsat-fusedapache-2.0
Downloads13M/mo
Accessopen · no token
Taskaudio-classification
Updated6mo ago
Downloads963k/mo
Accessopen · no token
Taskaudio-classification
Updated2.3y ago
🤗 audeering/wav2vec2-large-robust-12-ft-emotion-msp-dimcc-by-nc-sa-4.0
Downloads575k/mo
Accessopen · no token
Taskaudio-classification
Updated1.8y ago
🤗 MIT/ast-finetuned-audioset-10-10-0.4593bsd-3-clause
Downloads558k/mo
Accessopen · no token
Taskaudio-classification
Updated2.8y ago
🤗 m-a-p/MERT-v1-330Mcc-by-nc-4.0
Downloads491k/mo
Accessopen · no token
Taskaudio-classification
Updated14mo ago
To analyze music audio, extract measures, determine beats per minute (BPM), and perform similar tasks, I found several relevant models on Hugging Face that focus on audio classification. Below are five promising candidates:
Top Audio Classification Models for Music Analysis
| Model | Downloads | License | Gated | Updated | Usage |
|---|---|---|---|---|---|
| laion/clap-htsat-fused | 12,753,390 | Apache-2.0 | No | Jan 12, 2026 | General audio classification and feature extraction |
| alefiury/wav2vec2-large-xlsr-53-gender-recognition-librispeech | 962,532 | Apache-2.0 | No | Mar 23, 2024 | Gender recognition in audio |
| audeering/wav2vec2-large-robust-12-ft-emotion-msp-dim | 574,696 | CC BY-NC-SA 4.0 | No | Sep 19, 2024 | Emotion recognition in audio |
| MIT/ast-finetuned-audioset-10-10-0.4593 | 558,113 | BSD-3-Clause | No | Sep 6, 2023 | General audio classification |
| m-a-p/MERT-v1-330M | 491,396 | CC BY-NC 4.0 | No | May 25, 2025 | Music feature extraction |
Highlights
- Most Downloads: The CLAP model (
laion/clap-htsat-fused) is not only the most widely adopted model (over 12 million downloads) but also supports feature extraction and zero-shot classification, making it a versatile choice for diverse music-related tasks. - Emotion and Gender Recognition: Models like Wav2Vec2 versions specifically target emotion and gender detection, which could be beneficial if your analysis involves understanding the emotional context of music.
- Licensing: The majority are under permissive licenses (Apache-2.0, BSD-3-Clause) except for Audeering's model and MERT-v1, which have Creative Commons licenses that might come with usage restrictions.
Next Steps
- Each of these models can be integrated into your application for analyzing music audio profiles. The CLAP model is particularly recommended for comprehensive analysis due to its high adoption and feature extraction capabilities.
- Ensure to review the documentation of each model for details about implementation, as this will influence how you achieve the specific measures and analysis required for your project.
🤗 huggingface.co/laion/clap-htsat-fused🤗 huggingface.co/alefiury/wav2vec2-large-xlsr-53-gender-recognition-librispeech🤗 huggingface.co/audeering/wav2vec2-large-robust-12-ft-emotion-msp-dim🤗 huggingface.co/MIT/ast-finetuned-audioset-10-10-0.4593🤗 huggingface.co/m-a-p/MERT-v1-330M🤗 huggingface.co/facebook/mms-lid-256🤗 huggingface.co/OpenMuQ/MuQ-large-msd-iter🤗 huggingface.co/xbgoose/hubert-large-speech-emotion-recognition-russian-dusha-finetuned