YAMNet
Acoustics processing
Audio event classification
Package or Library
YAMNet is an AudioSet-trained audio event classifier (521 classes) that outputs class probabilities and embeddings.
| Model metadata | Value |
|---|---|
| Tool Type | Package or Library |
| Broad Task | Acoustics processing |
| Specific Task | Audio event classification |
| Model type | MobileNetV1-based CNN |
| Description | YAMNet is an AudioSet-trained audio event classifier (521 classes) that outputs class probabilities and embeddings. |
| Task Specific | Yes (embeddings for transfer learning) |
| Ecology Specific | No |
| Language(s) | Python, TensorFlow |
| Last Edited Time | 3/9/24 |
| Related Publication(s) | https://doi.org/10.1109/ICASSP.2017.7952261 |
| Dependencies | tensorflow, tensorflow-hub, numpy |
| Tool URL (Github etc.) | https://github.com/tensorflow/models/tree/master/research/audioset/yamnet |
| Last Update (time ago) | Last updated within 2 years |
| License | Apache-2.0 |
| Contact Name | |
| Contact Email | |
| Contact Responsiveness | |
| HuggingFace URL | |
| Reproducibility Method |