itisfoundation / jupyter-machine-learning

jupyter-ml-tensorflow and jupyter-ml-pytorch

This repository contains the source code for two o²S²PARC Services: jupyyter-ml-tensorflow and jupyter-ml-pytorch

Building the docker images:

make build

Test the built images locally:

pytorch

make run-pytorch-local

tensorflow

make run-tensorflow-local

Note that the validation directory will be mounted inside the service.

Raising the version can be achieved via one for three methods. The major,minor or patch can be bumped, for example:

pytorch

make version-pytorch-patch

tensorflow

make version-tensorflow-patch

If you already have a local copy of o²S²PARC running and wish to push data to the local registry:

make publish-local

jupyter-machine-learning's Issues

pytorch service fails to start when README.ipynb is missing

pytorch service fails to start when the state folder does not include README!

• Open pytorch service in osparc
• Delete work/README.ipynb
• close study
• reopen
• it fails
  

the log in the GUI cannot be read

the log in portainer reveals the real issue

Update service base image

The images fail to build in the CI with this error:
docker.io/nvidia/cuda:11.2.1-cudnn8-runtime-ubuntu18.04: not found

Because this image doesn't exist anymore: nvidia/cuda:11.2.1-cudnn8-runtime-ubuntu18.04 (see Dockerhub list here).

Should we go with nvidia/cuda:11.2.2-cudnn8-runtime @alessandrofasse?

Here's the table of compatibilities for tensorflow and tensorflow-gpu: https://www.tensorflow.org/install/source#gpu
Is there something similar for pytorch?

Cuda Version out of Sync with Tensorflow

Currently we have Cuda 11.0 including cudNN 8.0 is installed. However, for tensorflow 2.9.X we need to have 11.2 and 8.1 as shown on this page

https://www.tensorflow.org/install/source#gpu