At the moment, we are interpreting GET requests returned with status code 200 but with
as ZeroGPU quota errors. However, evidently this is also what is returned for other errors thrown within Hugging Face spaces. For instance, one of our spaces was throwing the following error:
Traceback (most recent call last):
File "/usr/local/lib/python3.10/site-packages/spaces/zero/wrappers.py", line 256, in run_task
res = task(*args, **kwargs) # pyright: ignore [reportCallIssue]
File "/home/user/app/app.py", line 236, in generate_accompaniment
f"({ops.max_time(melody_history, seconds=True):.2f}s of events captured)")
NameError: name 'melody_history' is not defined
This correspondingly generated the following popup on Hugging Face:
ZeroGPU worker error
NameError
However, in HARP the "ZeroGPU quota reached" error message was provided to the user. We need to find a way to differentiate between these different types of errors on Hugging Face.
At the moment, we are interpreting GET requests returned with status code 200 but with
as ZeroGPU quota errors. However, evidently this is also what is returned for other errors thrown within Hugging Face spaces. For instance, one of our spaces was throwing the following error:
This correspondingly generated the following popup on Hugging Face:
However, in HARP the "ZeroGPU quota reached" error message was provided to the user. We need to find a way to differentiate between these different types of errors on Hugging Face.