Notebook Images Overview

Describes notebook images available in HPE Ezmeral Unified Analytics Software and their uses.

Notebook images contain all the necessary software dependencies and configurations needed to run machine learning workflows. By using notebook images, you can collaborate, share, and deploy models with minimal compatibility issues.

Image Format

The images follow the following format:
<base-repository>/<image-name>:<image-tag>
For example:
gcr.io/mapr-252711/kubeflow/notebooks/jupyter-scipy:ezaf-fy23-q4-sp4-r9
Here,
  • base-repository: gcr.io/mapr-252711/kubeflow/notebooks
  • image-name: jupyter-scipy
  • image-tag: ezaf-fy23-q4-sp4-r9

Supported Notebook Images

The following table describes the notebook images available in HPE Ezmeral Unified Analytics Software and their uses.
Notebook Images Descriptions Uses
gcr.io/mapr-252711/kubeflow/notebooks/jupyter-scipy:<image-tag> This image is packaged with data science packages, including Pandas for data manipulation, Matplotlib and Bokeh for advanced plotting, and statistical tools such as SciPy and Statsmodels. Use this image to perform data analysis, manipulation, and visualization that doesn’t require machine learning libraries such as TensorFlow or PyTorch.
gcr.io/mapr-252711/kubeflow/notebooks/jupyter-pytorch-full:<image-tag> This image is packaged with data science packages and is integrated with PyTorch machine learning libraries for CPU-based tasks. This image does not have GPU acceleration capability. Use this image to perform data analysis, manipulation, and visualization for CPU-based machine learning tasks using PyTorch library.

gcr.io/mapr-252711/kubeflow/notebooks/jupyter-pytorch-cuda-full:<image-tag> This image is packaged with data science packages and is integrated with PyTorch machine learning libraries for GPU-based tasks. Use this image to perform data analysis, manipulation, and visualization for GPU-based machine learning tasks using PyTorch library for faster model training and data processing.
gcr.io/mapr-252711/kubeflow/notebooks/jupyter-tensorflow-full:<image-tag> This image is packaged with data science packages and is integrated with TensorFlow machine learning libraries for CPU-based tasks. This image does not have GPU acceleration capability. Use this image to perform data analysis, manipulation, and visualization for CPU-based machine learning tasks using TensorFlow library.
gcr.io/mapr-252711/kubeflow/notebooks/jupyter-tensorflow-cuda-full:<image-tag> This image is packaged with data science packages and is integrated with TensorFlow machine learning libraries for GPU-based tasks. Use this image to perform data analysis, manipulation, and visualization for GPU-based machine learning tasks using TensorFlow library for faster model training and data processing.
gcr.io/mapr-252711/kubeflow/notebooks/jupyter-data-science:<image-tag> This image is integrated with Tensorflow and PyTorch packages, including various other tools for data analysis, machine learning, and visualization. Use this image that is integrated with data science libraries to perform data science tasks requiring deep learning capabilities of TensorFlow and PyTorch.
gcr.io/mapr-252711/kubeflow/notebooks/codeserver:<image-tag>

This image enables you to run Visual Studio Code in the browser where you can edit and develop code in a remote server setup.

This image features a VS Code environment, providing a code-server that runs Visual Studio Code in the browser, allowing for rich code editing and development experience in a remote server setup.

In HPE Ezmeral Unified Analytics Software, the codeserver image includes VS Code and Python installation, and VS Code Python extension.

Use this image to run Visual Studio Code in the browser where you can edit and develop code in a remote server setup.

Image and Package Support

For a list of supported notebook images and included packages in HPE Ezmeral Unified Analytics Software, see Notebook Images.