You can use published Docker image as a base for your notebook execution. It can come either from a private or a public Docker registry, just like with any other Valohai execution.
You can also use any available image from hub.docker.com for example:
- https://hub.docker.com/r/jupyter/tensorflow-notebook/tags
- https://hub.docker.com/r/pytorch/pytorch/tags
The Docker image doesn't need to contain any Valohai specific packages. It's recommended that the image contains Jupyter Notebook or Jupyter Lab but if it doesn't Valohai will install it during runtime.
This allows you to use a wide variety of different images and environments.
You can also install packages directly inside the notebook by running a cell with !pip install:
!pip install matplotlib
Comments
0 comments
Please sign in to leave a comment.