| --- |
| license: mit |
| tags: |
| - audio |
| - source-separation |
| - drums |
| - onnx |
| pipeline_tag: audio-to-audio |
| --- |
| |
| # DrumSep ONNX |
|
|
| Drum element separation model converted to ONNX format for use with Splitzo. |
|
|
| ## Model Details |
|
|
| - **Original Author**: [inagoy/drumsep](https://github.com/inagoy/drumsep) |
| - **License**: MIT |
| - **Format**: ONNX |
| - **Size**: ~335 MB |
|
|
| ## Stems |
|
|
| Separates drum tracks into 4 elements: |
| 1. **Kick** (bombo) |
| 2. **Snare** (redoblante) |
| 3. **Cymbals** (platillos) |
| 4. **Toms** |
|
|
| ## Technical Specs |
|
|
| | Parameter | Value | |
| |-----------|-------| |
| | Sample Rate | 44100 Hz | |
| | Segment Length | 40 seconds (1,764,000 samples) | |
| | STFT n_fft | 4096 | |
| | STFT hop_length | 1024 | |
|
|
| ## Usage |
|
|
| This model is automatically downloaded by the Splitzo app on first use. |
|
|
| **Input**: Pre-separated drum stem (stereo, 44.1kHz) |
| **Output**: 4 separated drum elements |
|
|
| ## Attribution |
|
|
| This ONNX conversion is based on the original DrumSep model by [@inagoy](https://github.com/inagoy). |
|
|
| ``` |
| MIT License |
| |
| Copyright (c) 2023 inagoy |
| |
| Permission is hereby granted, free of charge, to any person obtaining a copy |
| of this software and associated documentation files (the "Software"), to deal |
| in the Software without restriction, including without limitation the rights |
| to use, copy, modify, merge, publish, distribute, sublicense, and/or sell |
| copies of the Software... |
| ``` |
|
|
| ## Conversion |
|
|
| Converted from PyTorch checkpoint using `tools/convert_drumsep_onnx.py` with STFT operations moved outside the model for ONNX compatibility. |
|
|
