504 Gateway Time-out in Inference Endpoints

On calling suno/bark model (which is a large model for TTS) ,it seems to give 504 Gateway Time-out on a POST request to the API.

Here’s the link to the repository :

Can you please take a look at it ?

Thank you.

I am having the same issue with the Facebook MusicGen small model on hosted inference. 504 timeouts for the last few hours.

I’ve been getting 504s for the last 8 hours or so. I’ve also received a couple 424s which is most certainly not the correct http code. Outside of that, the responses that do come back take much longer than usual.