Spaces:

fffiloni

/

image-to-sound-fx

like 101

Runtime error

Yes we are working on it, sometimes loading result fails

Refresh the page and try again and again until it works for you :)

@haoheliu do you mind considering to replace the gr.make_waveform() output by a gr.Audio() output ?

There is a bug with the waveform, causing the errors when trying to load the waveform's video from the audioLDM space

@fffiloni I'm afraid I'd prefer to keep the waveform visualization. Maybe you can also choose to display the video? Or there might be other workarounds.

I understand :)
I started by displaying the video, but we get the same error
see : https://huggingface.co/spaces/fffiloni/image-to-sound-fx-debug

What you can do instead to keep the waveform on your side, is adding a hidden gr.Audio() output to the output list
So then i can directly load the audio incognito, what do you think ?

I can do a PR for you

@haoheliu i made a PR for adding a hidden audio output
https://huggingface.co/spaces/haoheliu/audioldm-text-to-audio-generation/discussions/55

OK merged. Hope things can become easier for you.

Thank you 🙏 I’ll keep you updated

Forgot the brackets but just saw that you corrected it

https://huggingface.co/spaces/haoheliu/audioldm-text-to-audio-generation/blob/761f2df04b7845070716a5219ed9c00fe9047b55/app.py#L267

You should get rid of the brackets line 267

I see, will do

And at line 247 too, get rid of the brackets for the outputs value

There still seems to be an error. I'll temporally roll back. Please create a fork, check things are alright, and then make the PR. Thanks.

Okay 👍