Skip to content

A Mel-Band-Roformer model that is used to separate aspiration.

Notifications You must be signed in to change notification settings

SUC-DriverOld/Aspiration-Mel-Band-Roformer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Aspiration_Mel_Band_Roformer

A Mel-Band-Roformer model that is used to separate aspiration.

How to use

  • Clone this repository.
git clone https://github.com/SUC-DriverOld/Aspiration-Mel-Band-Roformer.git
  • Install the requirements.
pip install -r requirements.txt
usage: inference.py [-h] [--config_path CONFIG_PATH] [--model_path MODEL_PATH] [--input_folder INPUT_FOLDER] 
                    [--store_dir STORE_DIR] [--device_ids DEVICE_IDS [DEVICE_IDS ...]]

options:
  -h, --help                    show this help message and exit
  --config_path CONFIG_PATH     path to config yaml file
  --model_path MODEL_PATH       Location of the model
  --input_folder INPUT_FOLDER   folder with songs to process
  --store_dir STORE_DIR         path to store model outputs
  --device_ids DEVICE_IDS       list of gpu ids

Description

Huggingface model repository: https://huggingface.co/Sucial/Aspiration_Mel_Band_Roformer
You can try listening to the performance of this model here.

Instruments: aspiration, other
Dataset: My own datasets(171 songs for training and 17 songs for validation).
Finetuned from: model_mel_band_roformer_ep_3005_sdr_11.4360.ckpt
Configs: config_aspiration_mel_band_roformer.yaml

Model Name aspiration_mel_band_roformer_sdr_18.9845.ckpt aspiration_mel_band_roformer_less_aggr_sdr_18.1201.ckpt
Epoch 123 27
Instr SDR aspiration 9.8554 9.0704
Instr SDR other 28.1136 27.1699
SDR Avg 18.9845 18.1201

Thanks to

About

A Mel-Band-Roformer model that is used to separate aspiration.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages