Hello there, can we please get an update on the progress to diagnose and fix this issue? Standard models seem to function normally, but LoRas and finetunes give the error "failed to fetch the blob" or otherwise state that they are unavailable to be deployed via any of the available inference providers. My own spaces' report in the container logs that they are experiencing a mass JSONDecodeError and return the following error:

JSONDecodeError: Expecting value: line 1 column 1 (char 0)

I hope you've had some luck in identifying why only some of the text-to-image models broke.