HOW TO USE OTHER DATASET TO TRAIN Multi-Decoder-DPRNN MODEL

asteroid-team / asteroid

The PyTorch-based audio source separation toolkit for researchers

https://asteroid-team.github.io/

MIT License

2.21k stars 419 forks source link

HOW TO USE OTHER DATASET TO TRAIN Multi-Decoder-DPRNN MODEL #670

Open xiaoyu942 opened 1 year ago

xiaoyu942 commented 1 year ago

❓ Questions and Help

Hello, I would like to ask for advice. I want to use my own Chinese data to train a "Multi-Decoder-DPRNN" model, but I don't know what the data requirements are. Where can I see the structure of the data set? How should I prepare my data? Thanks!

mpariente commented 1 year ago

@JunzheJosephZhu, do you want to help ?

xiaoyu942 commented 1 year ago

I have a Chinese dataset and I would like to use it to train a model for speech separation， like your Multi-Decoder-DPRNN project

JunzheJosephZhu commented 1 year ago

Hi, The dataset python file is at https://github.com/asteroid-team/asteroid/blob/master/egs/wsj0-mix-var/Multi-Decoder-DPRNN/wsj0_mix_variable.py