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.
- beautifulsoup4
- bokeh
- cloudpickle
- cython
- dask
- dill
- distributed
- h5py
- ipywidgets
- matplotlib
- pandas
- plotly
- pyarrow
- s3fs
- scikit-image
- scikit-learn
- scipy
- seaborn
- sqlalchemy
- statsmodels
- torch
- torchvision
- jupyter-bokeh
- jupyter-tensorboard
A pre-built version of the pytorch notebook based on Thoth s2i-minimal-notebook, can be found at quay.io:
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
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-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-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