Roformer Models
BS-RoFormer & MelBand RoFormer — State-of-the-art music source separation
Model source
- Kind
- Other
- File size
- 503 MB–3.51 GB
- Version
- vRevision d323194
- Updated
- Oct 11, 2026
- Downloads
- 0
- Source
- Hugging Face
Source description
RoFormer Stem Separation Models (Safetensors)
BS-RoFormer & MelBand RoFormer — State-of-the-art music source separation
Pretrained weights converted to safetensors format for use with Mæstræa AI Workstation.
Models
BS-RoFormer (Band-Split RoPE Transformer)
| Variant | SDR | Task | Path |
|---|---|---|---|
| Vocals (viperx) | 12.97 | Vocal/instrumental separation | bs_roformer/vocals_viperx/ |
| Multi-stem | 9.65 | 4-stem (bass/drums/vocals/other) | bs_roformer/multistem/ |
MelBand RoFormer (Mel-Band RoPE Transformer)
| Variant | SDR | Task | Path |
|---|---|---|---|
| Vocals (KimberleyJensen) | 10.98 | Best vocal isolation | mel_band_roformer/vocals_kj/ |
| Vocals (viperx) | 11.43 | Vocal/instrumental separation | mel_band_roformer/vocals_viperx/ |
| Dereverb (anvuew) | 19.17 | Remove reverb from audio | mel_band_roformer/dereverb/ |
| Denoise (aufr33) | 27.99 | Remove noise from audio | mel_band_roformer/denoise/ |
Architecture
Both models use the Band-Split RoPE Transformer architecture from lucidrains/BS-RoFormer:
- BS-RoFormer: Splits spectrogram into uniform-width subbands
- MelBand RoFormer: Splits using mel-scale (perceptually-weighted) overlapping bands
Both significantly outperform HTDemucs on vocal separation tasks.
Usage
Each model directory contains:
model.safetensors— Model weightsconfig.yaml— Architecture configuration (required for model instantiation)
Requires bs-roformer Python package: pip install bs-roformer
Credits
- Architecture: lucidrains/BS-RoFormer
- Training framework: ZFTurbo/Music-Source-Separation-Training
- BS-RoFormer vocals: viperx via TRvlvr
- MelBand vocals: KimberleyJensen, viperx
- MelBand dereverb: anvuew
- MelBand denoise: aufr33
- Conversion & Mirror by: AEmotionStudio
License
MIT — same as all upstream model releases.
Sources
1 source
Sources
1 sourceVerified Aug 14
Model artifacts
11 artifacts
Model artifacts
11 artifactsbs_roformer/multistem/model.safetensors
safetensors · 503 MB · SHA-256 14c9079fa428…9098 · Hugging Face
Downloadbs_roformer/vocals_viperx/model.safetensors
safetensors · 610 MB · SHA-256 fa2965772061…5a8a · Hugging Face
Downloadmel_band_roformer/4stems_large/model.safetensors
safetensors · 3.51 GB · SHA-256 3762f36cd3bb…c5b9 · Hugging Face
Downloadmel_band_roformer/denoise/model.safetensors
safetensors · 871 MB · SHA-256 721e2b13e30e…bb93 · Hugging Face
Downloadmel_band_roformer/dereverb/model.safetensors
safetensors · 871 MB · SHA-256 20af72ee52ab…a75c · Hugging Face
Downloadmel_band_roformer/inst_becruily/model.safetensors
safetensors · 871 MB · SHA-256 2838a953b925…fdbb · Hugging Face
Downloadmel_band_roformer/instvoc_duality/model.safetensors
safetensors · 1.60 GB · SHA-256 25a4a6e78189…fd75 · Hugging Face
Downloadmel_band_roformer/vocals_becruily/model.safetensors
safetensors · 871 MB · SHA-256 89e4c8ed4b40…55d9 · Hugging Face
Downloadmel_band_roformer/vocals_kj/model.safetensors
safetensors · 871 MB · SHA-256 03bfc70b43df…3095 · Hugging Face
Downloadmel_band_roformer/vocals_unwa_ft/model.safetensors
safetensors · 871 MB · SHA-256 35986ae128e2…fbb2 · Hugging Face
Downloadmel_band_roformer/vocals_viperx/model.safetensors
safetensors · 961 MB · SHA-256 505a93615bac…8937 · Hugging Face
DownloadSource excerpts
2 excerpts
Source excerpts
2 excerpts--- license: mit tags: - audio - music - source-separation - stem-separation - roformer - safetensors - maestraea pipeline_tag: audio-to-audio --- # RoFormer Stem Separation Models (Safetensors) **BS-RoFormer & MelBand RoFormer — State-of-the-art music source separation** > Pretrained weights converted to safetensors format for use with [Mæstræa AI Workstation](https://github.com/AEmotionStudio/Maestraea). ## Models ### BS-RoFormer (Band-Split RoPE Transformer) | Variant | SDR | Task | Path | |---------|-----|------|------| | Vocals (viperx) | 12.97 | Vocal/instrumental separation | `bs_roformer/vocals_viperx/` | | Multi-stem | 9.65 | 4-stem (bass/drums/vocals/other) | `bs_roformer/multistem/` | ### MelBand RoFormer (Mel-Band RoPE Transformer) | Variant | SDR | Task | Path | |---------|-----|------|------| | Vocals (KimberleyJensen) | 10.98 | Best vocal isolation | `mel_band_roformer/vocals_kj/` | | Vocals (viperx) | 11.43 | Vocal/instrumental separation | `mel_band_roformer/vocals_viperx/` | | Dereverb (anvuew) | 19.17 | Remove reverb from audio | `mel_band_roformer/dereverb/` | | Denoise (aufr33) | 27.99 | Remove noise from audio | `mel_band_roformer/denoise/` | ## Architecture Both models use the Band-Split RoPE Transformer architecture from [lucidrains/BS-RoFormer](https://github.com/lucidrains/BS-RoFormer): - **BS-RoFormer**: Splits spectrogram into uniform-width subbands - **MelBand RoFormer**: Splits using mel-scale (perceptually-weighted) overlapping bands Both significantly outperform HTDemucs on vocal separation tasks. ## Usage Each model directory contains: - `model.safetensors` — Model weights - `config.yaml` — Architecture configuration (required for model instantiation) Requires `bs-roformer` Python package: `pip install bs-roformer` ## Credits - **Architecture**: [lucidrains/BS-RoFormer](https://github.com/lucidrains/BS-RoFormer) - **Training framework**: [ZFTurbo/Music-Source-Separation-Training](https://github.com/ZFTurbo/Music-Source-Separation-Training) - **BS-RoFormer vocals**: [viperx](https://github.com/playdasegunda) via [TRvlvr](https://github.com/TRvlvr/model_repo) - **MelBand vocals**: [KimberleyJensen](https://github.com/KimberleyJensen), [viperx](https://github.com/playdasegunda) - **MelBand dereverb**: [anvuew](https://github.com/anvuew) - **MelBand denoise**: [aufr33](https://github.com/aufr33) - **Con...
Source context: 0 downloads · 3 likes · Pipeline audio-to-audio · Repo AEmotionStudio/roformer-models
