[jupyter/tensorflow-notebook] Kernel stays in Busy mode indefinitely

[ravigunasekara]

[jupyter/tensorflow-notebook] Kernel stays in Busy mode indefinitely

[benz0li]

@ravigunasekara Bind mounting does not work with VirtioFS: docker/for-mac#6243 (comment)

Use gRPC FUSE.

[mathbunnyru]

We only push latest images to Quay.io, could you please try using quay.io/jupyter/tensorflow-notebook image?

[mathbunnyru]

Also, please, try executing your command without volume mounted - do you still have an error?

[ravigunasekara]

Thanks for the response Ayas, with this image, it straightaway kills the kernel and restarts. when I try to import tensorflow after restart it kills again and restarts. Forgot to mention am using a MAC with M1 chip. I have tried it on my old Mac which has Intel chipset and it works there. Am good for now but leaving this open as this could be the case for other mac users with an M1 chip. Let me know if you need further details for investigations. thanks

[mathbunnyru]

docker run --pull always -p 8888:8888 quay.io/jupyter/tensorflow-notebook + import tensorflow works fine for me on M2 MacBook Air, Docker Desktop 4.26.1, macOS Sonoma 14.2.1.
I remember a case when upgrading to Sonoma resolved the issue (it was about emulating x86_64 on a apple-silicon Mac though): #1999 (comment)

Could you please try to close all the jupyterlab tabs in a browser, run the command above, and then try to import tensorflow in a private tab?
Please, attach your logs from the terminal as well.
Also, please, make sure, you have the same Docker Desktop version.

[benz0li]

Out of curiosity: Does it work with my JupyterLab Python docker stack?

On your Apple computer

docker run -it --rm \
  -p 8888:8888 \
  -v "${PWD}":/home/jovyan \
  glcr.b-data.ch/jupyterlab/python/base

In the container

JupyterLab Terminal

pip install tensorflow

python

Python REPL

import tensorflow as tf

---

For more information, see https://github.com/b-data/jupyterlab-python-docker-stack.
(with b-data's/my docker stack one is supposed to mount the entire home directory).

1. Multi-arch: linux/amd64, linux/arm64/v8
   ℹ️ Runs on Apple M series using Docker Desktop.
2. Base image: Debian instead of Ubuntu
   ℹ️ CUDA-enabled images are Ubuntu-based.
3. IDE: code-server next to JupyterLab
   ℹ️ code-server = VS Code in the browser.
4. Just Python – no Conda / Mamba