Zero GPU Worker Error

Spaces
21

Thank you!
Well, I think we understand the situation (in terms of the program). I’ll leave the rest to you…:sleepy_face:

1 Like
22

image
image1773×380 35.6 KB

got this in merterbak spaces here Gpt Oss 20b Demo - a Hugging Face Space by merterbak

and mine : Med-GPT-OSS-20B - a Hugging Face Space by Tonic same story

factory restart didnt help , spaces working fine beforehand

seems like a server-side partial outtage to me

23

Duplicating or restarting seems to work. That Space does not contain any code that causes problems with PyTorch 2.8 (as far as I can see).

ponyrealismreboot
ponyrealismreboot1130×539 134 KB

24

With that code, restarting the space should probably do the trick.

25

well , you were correct.

sometimes i just need to talk to the king of the community help (that’s you) , change nothing , and simply your attention to the matter resolves all the bugs.

so all my thanks for your time :saluting_face::folded_hands:t2: much appreciated

1 Like
26

Always the same problem for me. But I probably didn’t understand what I should do ….

1 Like
27

OP here. I didn’t get much of the programming lingo (noob here!), but the problems resolved today. Thanks, everyone!! Much appreciated.

1 Like
28

Did you do something?

1 Like
29

Ive been having the same problem with TRELLIS. TRELLIS - a Hugging Face Space by trellis-community It only just started and I have my subscription up.

1 Like
30

For Pro subscribers, duplicating the Space will usually fix the issue.
If manual adjustments are necessary (like WAN 2.2 space above, etc.), you will need to modify the code, but if the issue can be resolved by simply restarting (or rebuilding), duplicating the Space will suffice.

31

It worked after duplicating. No code changes seem to be necessary. If the author just restarts the space, the original space should be fixed.

trellisrst
trellisrst1299×593 148 KB

32

Thanks! Strange wave of this one error. Hope it doesn’t happen again

1 Like
33

A few minutes ago, several Zero GPU spaces suddenly crashed, so I tried rebuilding them, and it seems that the PyTorch version has become a little more flexible. This is probably an intentional fix…

===== Build Queued at 2025-08-25 08:37:20 / Commit SHA: e548b67 =====

--> FROM docker.io/library/python:3.10.13@sha256:d5b1fbbc00fd3b55620a9314222498bebf09c4bf606425bf464709ed6a79f202
DONE 0.0s

--> RUN apt-get update && apt-get install -y 	git 	git-lfs 	ffmpeg 	libsm6 	libxext6 	cmake 	rsync 	libgl1 	&& rm -rf /var/lib/apt/lists/* 	&& git lfs install
CACHED

--> RUN pip install --no-cache-dir pip -U && 	pip install --no-cache-dir 	datasets 	"huggingface-hub>=0.19" "hf_xet>=1.0.0,<2.0.0" "hf-transfer>=0.1.4" "protobuf<4" "click<8.1" "pydantic~=1.0" torch==2.4.0
CACHED

--> RUN apt-get update && apt-get install -y fakeroot &&     mv /usr/bin/apt-get /usr/bin/.apt-get &&     echo '#!/usr/bin/env sh\nfakeroot /usr/bin/.apt-get $@' > /usr/bin/apt-get &&     chmod +x /usr/bin/apt-get && 	rm -rf /var/lib/apt/lists/* && 	useradd -m -u 1000 user
CACHED

--> WORKDIR /home/user/app
CACHED

--> RUN wget --progress=dot:giga https://developer.download.nvidia.com/compute/cuda/12.9.0/local_installers/cuda_12.9.0_575.51.03_linux.run -O cuda-install.run 	&& fakeroot sh cuda-install.run --silent --toolkit --override 	&& rm cuda-install.run
CACHED

--> COPY --chown=1000:1000 --from=root / /
CACHED

--> RUN 	apt-get update && 	apt-get install -y curl && 	curl -fsSL https://deb.nodesource.com/setup_20.x | bash - && 	apt-get install -y nodejs && 	rm -rf /var/lib/apt/lists/* && apt-get clean
CACHED

--> Restoring cache
DONE 198.8s