search

tsurumeso / vocal-remover

Vocal Remover using Deep Neural Networks
MIT License
1.47k stars 215 forks source link

Enriching trained data model #150

Open dangtaikhtn opened 9 months ago

dangtaikhtn commented 9 months ago

Hi @tsurumeso

Firstly I would like to say thank you to your awesome things that you shared. And I have a question that am I able to use the trained data model to append some new songs instead of training a new model from scratch. Hopefully you can see my message and let me know if you have some suggestions.

tsurumeso commented 9 months ago

Thank you for your kind words. You can use the --pretrained_model option like this:

python train.py --dataset path/to/dataset --pretrained_model path/to/pretrained/model --reduction_rate 0.5 --mixup_rate 0.5 --gpu 0