metadata
tags:
- coreml
- audio
- music-source-separation
- schism
license: mit
DeepFilterNet3 Enhancement
CoreML model for Schism audio separation framework.
DeepFilterNet3 speech enhancement (~2.3M params) — ERB spectral masking + deep filtering with learned complex FIR filters
Details
| Property | Value |
|---|---|
| Expert IDs | deepfilternet |
| Tasks | audioEnhancement |
| Quality | medium |
| Est. RTF | 0.1 |
| Peak Memory | 50 MB |
Usage
Install to your local Schism model store:
schism models install deepfilternet3
Or manually place in ~/.schism/models/deepfilternet3/v1.0.0/.