leimao / singing-voice-separation-rnn Goto Github PK

I can't download data set by the link no more because the link is unavailable.

So I got the data set by another link but I still couldn't run the code.

The error says that 'FileNotFoundError: [Errno 2] No such file or directory: 'data/MIR1K/train.txt'.

I have downloaded the dataset with many different links, but train.txt is not in any folder.

What is the content of train.txt? Please let me know...

Dear Friend.

I want to ask a question from you. !!
in model, why is this necessary?

assert len(num_hidden_units) == num_rnn_layer

and why can't we change the number of layers with this exception?

Please guide..

Greetings.

Dear friend:

     If I input a mono MIR-1K audio, it can be tested and how to eval?  

please guide...

Dear friend.
I have one question. When I use this code to deal with Singing and voice mixture, it works well. But when I use it to process speech mixture, it seems to have no effect. Can you guide why？

Regards.

Dear friend.

I have one more question.
no matter how much the model is trained. the loss function does not go below 1.2 or we can say 1.0.

Can you guide why? or this is multiplied by any of the factor?

Because as far as I understand, for good training, the loss should decrease 0.1 or at least nearby.

If you can guide because I am not so expert and trying to learn.

Greetings.