How can I split the fma-small dataset with genres?

Question

How can I split the fma-small dataset with genres?

DennisShaw opened this issue 7 years ago · comments

I can not find a way to split the source audio zip to 8 genres with 1000 tracks.I can't find a file to help me do this.Would you mind helping me? Please...Thanks

Michaël Defferrard · Answer 1 · Wed Feb 21 2018 06:05:46 GMT+0800 (China Standard Time)

You should download the fma_metadata.zip file which contains all the metadata. Then look at the usage notebook for guidance to manipulate that metadata.

DennisShaw · Answer 2 · Mon Feb 26 2018 12:54:44 GMT+0800 (China Standard Time)

But the fifth part of usage.ipynb( Genre classification) is not finished yet.

Michaël Defferrard · Answer 3 · Mon Feb 26 2018 19:31:38 GMT+0800 (China Standard Time)

You can look at the baselines notebook for more code about genre classification.

Abhinavram Ananth · Answer 4 · Tue Oct 16 2018 22:06:50 GMT+0800 (China Standard Time)

Hey, were

I can not find a way to split the source audio zip to 8 genres with 1000 tracks.I can't find a file to help me do this.Would you mind helping me? Please...Thanks

Hey, were you finally able to split the audios to 8 genres? If so, could you please help me out do the same?

Scott H. Hawley · Answer 5 · Thu Aug 22 2019 14:23:07 GMT+0800 (China Standard Time)

I posted a gist to do this at
https://gist.github.com/drscotthawley/eb4ffb1ec4de29632403c1db396e419a