Intel's Machine Learning using oneAPI - II

por Bernardo Bernardino Gameiro (Instituto de Física Corpuscular (IFIC))

miércoles 29 ene. 2025 14:00 → 18:00 Europe/Madrid

1001-Baja-B-3 - Paterna. Sala de Juntas (Universe)

Join us for an introductory exploration of Intel's Machine Learning Using oneAPI!

This workshop offers will consist in a series of practical examples showcasing how you can leverage oneAPI, an open standard and unified cross-architecture API, in your machine learning codes.
This facilitates the adoption of  heterogeneous computing, seamlessly targeting both CPU and GPU architectures with a single version. Attendees will appreciate the opportunity to adopt a vendor-independent approach, based on industry-wide standards.

In an era where computational advancements are paramount in physics research, familiarity with SYCL presents a valuable asset. Join institutions such as CERN and Argonne National Laboratory by learning how to leverage SYCL's cutting-edge capabilities for scientific endeavors.

Participants will receive a certificate of attendance and access to the Intel's Developer Cloud.

The workshop is going to take place:

• Afternoon session
  • 14:00
  • Sala de Juntas
  • Complementary session for https://indico.ific.uv.es/e/oneapi_ml_202412

Participants are requested to create a free Intel Developer Cloud account prior to the workshop.

Class materials

• Machine Learning and Data Analysis for Nuclear Physics, European ERASMUS+ Master of Science program
• ML_workshop202501.ipynb
• ML_workshop202501.pdf

Motivation slides Motivation_OneAPI_BGameiro.pdf

Setup DevCloud_setup_new_users.pdf Jupyter Notebooks

Slides

• 01 MLoneAPI Course and oneAPI Introduction 20240920.pdf
• 02 Intel Extensions for scikit-learn Introduction 20240718.pdf
• 03 Intel Extensions for scikit-learn Advanced dpctl GPU 20240718.pdf
• 04 Numpy powered by oneAPI 20240718.pdf
• 05_Using_Intel_AI_Kit_and_XGBoost_For_Predictive_Modeling 20240718.pdf
• oneAPI Samples