search

andabi / music-source-separation

Deep neural networks for separating singing voice from music written in TensorFlow
795 stars 150 forks source link

No ikala file #32

Open ucasiggcas opened 5 years ago

ucasiggcas commented 5 years ago

Dear,

Thanks for your great project,

but in the config.py I find DATA_PATH = 'dataset/train/ikala'

so where is the train dataset?

With your mir-1k file how to change your code?

How to reproduce your project ?

Thanks a lot

Lychee Tsu

ieee-msn2017 commented 5 years ago

I have the same question. The iKala dataset is no longer available according to http://mac.citi.sinica.edu.tw/ikala/index.html. How can I use other dataset?

ucasiggcas commented 5 years ago

I have the same question. The iKala dataset is no longer available according to http://mac.citi.sinica.edu.tw/ikala/index.html. How can I use other dataset?

Did you get the dataset ? Could U share ?

Have you run success ? Please help me.

Thx

Tortoise17 commented 5 years ago

I think that dataset is no more in public domain. Its industrial based now.

ucasiggcas commented 5 years ago

@Tortoise17 Did you run the project success ?Could you reproduce it ? How to change the config ?

Tortoise17 commented 5 years ago

HI.. not perfectly. I will update you shortly.

ucasiggcas commented 5 years ago

WHo can Help me ? ANy one ? ANybody ?