S2I PyTorch Notebook

Custom Notebook built with Thoth s2i-minimal-notebook.

This custom notebook contains PyTorch related packages install with minimal notebook for Data Science usage. This image can be directly used if user needed these package while using the minimal jupyter notebook.

We have configured this repository to used pipenv and micropipenv python dependency managers.

List of packages in pytorch-notebook

- bokeh
- cython
- dask
- dill
- distributed
- h5py
- ipywidgets
- matplotlib
- pandas
- plotly
- pyarrow
- s3fs
- scikit-image
- scikit-learn
- scipy
- seaborn
- sqlalchemy
- statsmodels
- torch
- torchvision

List of extensions

- jupyter-bokeh
- jupyter-tensorboard
- jupyterlab-git
- jupyterlab-s3-browser
- elyra-python-editor-extension

Importing the PyTorch Notebook

A pre-built version of the pytorch notebook based on Thoth s2i-minimal-notebook, can be found at quay.io:

https://quay.io/repository/thoth-station/s2i-pytorch-notebook

This image could be imported into an OpenShift cluster using OpenShift ImageStream:

apiVersion: image.openshift.io/v1
kind: ImageStream
metadata:
  # (Below label is needed for Opendatahub.io/JupyterHub)
  # labels:
  #   opendatahub.io/notebook-image: "true"
  name: s2i-pytorch-notebook
spec:
  lookupPolicy:
    local: true
  tags:
  - name: latest
    from:
      kind: DockerImage
      name: quay.io/thoth-station/s2i-pytorch-notebook:latest

Building the PyTorch Minimal Notebook

Instead of using the pre-built version of the minimal notebook, you can build the minimal notebook from source code.

With Thoth advise

s2i build . quay.io/thoth-station/s2i-minimal-py38-notebook:latest \
--env ENABLE_PIPENV=1 \
--env THOTH_ADVISE=1 \
--env THOTH_DRY_RUN=0 \
--env THOTH_PROVENANCE_CHECK=1 \
s2i-pytorch-notebook

Without Thoth advise

s2i build . quay.io/thoth-station/s2i-minimal-py38-notebook:latest \
--env ENABLE_PIPENV=1 \
--env THOTH_ADVISE=0 \
--env THOTH_ERROR_FALLBACK=1 \
--env THOTH_DRY_RUN=1 \
--env THOTH_PROVENANCE_CHECK=0 \
s2i-pytorch-notebook