https://github.com/egorsmkv/vocos

https://github.com/egorsmkv/vocos

Last synced: 26 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/egorsmkv/vocos
• Owner: egorsmkv
• License: mit
• Created: 2023-06-16T20:00:29.000Z (over 1 year ago)
• Default Branch: master
• Last Pushed: 2023-06-25T12:22:42.000Z (over 1 year ago)
• Last Synced: 2024-01-09T23:25:33.861Z (10 months ago)
• Language: Python
• Size: 471 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Vocos

## Training

Prepare a filelist of audio files for the training and validation set:

```bash

find $TRAIN_DATASET_DIR -name *.wav > filelist.train

find $VAL_DATASET_DIR -name *.wav > filelist.val

```

Fill a config file, e.g. [vocos.yaml](configs%2Fvocos.yaml), with your filelist paths and start training with:

```bash

conda create -n vocos python=3.10

conda activate vocos

pip install -r requirements-all.txt

# train from the ground

python train.py -c configs/vocos.yaml

# using a checkpoint

python train.py -c configs/vocos.yaml --trainer.resume_from_checkpoint /home/yehor/Work/github/vocos/logs/lightning_logs/version_1/checkpoints/vocos_checkpoint_epoch=337_step=214332_val_loss=10.1883.ckpt

# run tensorboard to see the metrics

tensorboard --logdir logs/lightning_logs/version_1/

conda activate my

conda remove -n vocos --all

```