oneAPI: AI Analytics Toolkit [beartooth]

1 Overview | 2 Using | 2.1 Example:

Overview

  • AI Kit: The AI Kit gives data scientists, AI developers, and researchers familiar Python tools and frameworks to accelerate end-to-end data science and analytics pipelines on Intel architecture. The components are built using oneAPI libraries for low-level compute optimizations.

  • Using this toolkit, you can:

    • Deliver high-performance, deep learning training on Intel® XPUs and integrate fast inference into your AI development workflow with Intel-optimized, deep learning frameworks for TensorFlow and PyTorch, pretrained models, and low-precision tools. 

    • Achieve drop-in acceleration for data preprocessing and machine learning workflows with compute-intensive Python packages, Modin, scikit-learn, and XGBoost, optimized for Intel.

Using

These come in the form of conda environments. The following are currently available:

Version

Cluster

Conda Activate

Version

Cluster

Conda Activate

2023.2

Loren

/apps/u/opt/compilers/oneapi/2023.2/tensorflow/2.13.0 /apps/u/opt/compilers/oneapi/2023.2/pytorch/2.0.1.0 /apps/u/opt/compilers/oneapi/2023.2/modin/0.23.0.0

2023.1

Beartooth

/apps/u/opt/compilers/oneapi/2023.1/tensorflow/2.10.0.0 /apps/u/opt/compilers/oneapi/2023.1/pytorch/1.13.0.0 /apps/u/opt/compilers/oneapi/2023.1/modin/0.17.0.1

2022.3

Beartooth

/apps/u/opt/compilers/oneapi/2022.3/tensorflow/2.9.1.0 /apps/u/opt/compilers/oneapi/2022.3/pytorch/1.12.0 /apps/u/opt/compilers/oneapi/2022.3/modin/0.13.3

Example: