Repository for training models for music source separation. Repository is based on kuielab code for SDX23 challenge. The main idea of this repository is to create training code, which is easy to modify for experiments. Brought to you by MVSep.com.
Model can be chosen with --model_type
arg.
Available models for training:
- MDX23C based on KUIELab TFC TDF v3 architecture. Key:
mdx23c
. - Demucs4HT [Paper]. Key:
htdemucs
. - VitLarge23 based on Segmentation Models Pytorch. Key:
segm_models
. - TorchSeg based on TorchSeg module. Key:
torchseg
. - Band Split RoFormer [Paper, Repository] . Key:
bs_roformer
. - Mel-Band RoFormer [Paper, Repository]. Key:
mel_band_roformer
. - Swin Upernet [Paper] Key:
swin_upernet
. - BandIt Plus [Paper, Repository] Key:
bandit
. - SCNet [Paper, Official Repository, Unofficial Repository] Key:
scnet
. - BandIt v2 [Paper, Repository] Key:
bandit_v2
. - Apollo [Paper, Repository] Key:
apollo
. - TS BSMamba2 [Paper, Repository] Key:
bs_mamba2
.
- Note 1: For
segm_models
there are many different encoders is possible. Look here. - Note 2: Thanks to @lucidrains for recreating the RoFormer models based on papers.
- Note 3: For
torchseg
gives access to more than 800 encoders fromtimm
module. It's similar tosegm_models
.
To train model you need to:
- Choose model type with option
--model_type
, including:mdx23c
,htdemucs
,segm_models
,mel_band_roformer
,bs_roformer
. - Choose location of config for model
--config_path
<config path>
. You can find examples of configs in configs folder. Prefixesconfig_musdb18_
are examples for MUSDB18 dataset. - If you have a check-point from the same model or from another similar model you can use it with option:
--start_check_point
<weights path>
- Choose path where to store results of training
--results_path
<results folder path>
python train.py \
--model_type mel_band_roformer \
--config_path configs/config_mel_band_roformer_vocals.yaml \
--start_check_point results/model.ckpt \
--results_path results/ \
--data_path 'datasets/dataset1' 'datasets/dataset2' \
--valid_path datasets/musdb18hq/test \
--num_workers 4 \
--device_ids 0
All training parameters are here.
python inference.py \
--model_type mdx23c \
--config_path configs/config_mdx23c_musdb18.yaml \
--start_check_point results/last_mdx23c.ckpt \
--input_folder input/wavs/ \
--store_dir separation_results/
All inference parameters are here.
- All batch sizes in config are adjusted to use with single NVIDIA A6000 48GB. If you have less memory please adjust correspodningly in model config
training.batch_size
andtraining.gradient_accumulation_steps
. - It's usually always better to start with old weights even if shapes not fully match. Code supports loading weights for not fully same models (but it must have the same architecture). Training will be much faster.
configs/config_*.yaml
- configuration files for modelsmodels/*
- set of available models for training and inferencedataset.py
- dataset which creates new samples for traininggui-wx.py
- GUI interface for codeinference.py
- process folder with music files and separate themtrain.py
- main training codetrain_accelerate.py
- experimental training code to use withaccelerate
module. Speed up for MultiGPU.utils.py
- common functions used by train/validvalid.py
- validation of model with metricsensemble.py
- useful script to ensemble results of different models to make results better (see docs).
Look here: List of Pre-trained models
If you trained some good models, please, share them. You can post config and model weights in this issue.
Look here: Dataset types
Look here: Augmentations
Look here: GUI documentation or see tutorial on Youtube
@misc{solovyev2023benchmarks,
title={Benchmarks and leaderboards for sound demixing tasks},
author={Roman Solovyev and Alexander Stempkovskiy and Tatiana Habruseva},
year={2023},
eprint={2305.07489},
archivePrefix={arXiv},
primaryClass={cs.SD}
}