it's really not easy to reproduce the project ！

andabi / music-source-separation

Deep neural networks for separating singing voice from music written in TensorFlow

795 stars 150 forks source link

it's really not easy to reproduce the project ！ #39

Closed ucasiggcas closed 3 years ago

ucasiggcas commented 5 years ago

Dear,

You project maybe good, but no one has reproduced, 1.No ikala data 2.The data sets has no split into train/validataion/eval 3.No instruction about the config.py and how to modify it 4.No steps about the usage, I mean detail.

Thx

ucasiggcas commented 5 years ago

Who has got success please help me . Who can help me will be grateful. Who care me ?

Tortoise17 commented 5 years ago

HI @ ucasiggcas Did you get the iKala data?

ucasiggcas commented 5 years ago

HI @ ucasiggcas Did you get the iKala data?

No, please use the mir-1k

Tortoise17 commented 5 years ago

Send me your email.

ucasiggcas commented 5 years ago

Send me your email.

ucas010@126.com

cheriylan commented 3 years ago

Send me your email.

could you have ikala,can you send it to me

ucasiggcas commented 3 years ago

@cheriylan No Now I have got success ,but the rp is not perfect for complex situation Will close the issue Bye