Spaces:

hilamanor

/

audioEditing

like 264

Running on Zero

Hi, this is not false advertising, we just weren't clear enough and I'm sorry for that.
Duplicating the space copies the code as is, and thus it still applies automatic trimming.
The idea behind duplicating is that you can remove the trimming line in your own space (We employed this limit to enable multiple users to not wait in the queue too long):

In utils.py#L19 ,
just change
duration = min(audioldm.utils.get_duration(audio_path), 30)
to
duration = audioldm.utils.get_duration(audio_path)

Please update me if this does not solve your problem. I will add documentation in the code so that it will be clearer for the next person, thank you.

By the way, I made a duplicate, and I did change "duration = min(audioldm.utils.get_duration(audio_path), 30)" to "duration = audioldm.utils.get_duration(audio_path)", but I got an error called "connection error out". Help me.

Did you clone it locally or duplicate it still in the hugging face spaces website?
Can you provide the full error? If you did the latter, the error is available in the "logs" button when you try to view the space.
I ran into something similar 2 days ago where huggingface.co was temporarily unavailable, and so the pre-trained models couldn't be downloaded, and it threw a connection error (I attached image of how the error looked like).
However, it was just temporary (the models only need to be downloaded once..) and when the hugging face hub was back up again it worked (e.g. I just duplicated it now without problems).
If the error seems the same maybe try restarting the space so it will try to download the models again.

I'm closing this issue since I got no answer. Please write to me again if you have any troubles.
@jcksn @RuneLightLovely