VO OLA DEEP Training Facilities

Permalink:

https://documents.egi.eu/document/3747

Document #:

EGI-doc-3747-v2

Document type:

Other

Submitted by:

Giuseppe La Rocca

Updated by:

Giuseppe La Rocca

Document Created:

20 May 2021, 18:12

Contents Revised:

11 Jul 2023, 12:02

Metadata Revised:

11 Jul 2023, 12:02

Abstract:

The DEEP Training Facility is a distributed training facility for Machine Learning, Artificial Intelligence and Deep Learning models.
The DEEP Training Facility provides tools for building training, testing and training Machine Learning, Artificial Intelligence and Deep Learning models over distributed e-Infrastructures.
Models can be built from scratch or form existing and pre-trained models (transfer learning or model reuse). Ready to use models are available for transfer learning or reuse.

Features:
- Transparent training over distributed e-Infrastructures with GPU access.
- Docker based for model portability and reusability.
- Easy model integration with standards-based REST APIs.
- CLI and web user interface to interact with the system.
- OpenID Connect based identity.

Files in Document:

• EGI VO OLA DEEP Training Facility IFCA-LCG2_FINAL.pdf (205.7 kB)
• EGI VO OLA DEEP Training Facility NCG-INGRID-PT_FINAL.pdf (206.4 kB)

Other Files:

• EGI VO Deep trining facility SLA report 2021-05 - 2021...pdf (158.0 kB)
• EGI VO Deep trining facility SLA report 2021-11 - 2022...pdf (163.5 kB)
• EGI VO Deep trining facility SLA report 2022-05 - 2022...pdf (168.9 kB)
• EGI VO Deep trining facility SLA report 2022-11 - 2023...pdf (165.2 kB)
• EGI VO Deep trining facility SLA report 2023-05 - 2023...pdf (160.4 kB)

Topics:

• EGI.eu
• Activities:Operations

Authors:

• Giuseppe La Rocca