Skip to content

Releases: intel/ai-reference-models

Intel® AI Reference Models v3.2

25 Jul 14:30
Compare
Choose a tag to compare

Supported Frameworks

New features

  • Added new workloads scripts for the best known configurations on CPU platforms (Intel® Core™ Processors, Intel® Xeon® processors, Intel® Xeon® Scalable Processors) and GPU platforms (Intel® Data Center GPU Flex Series and Intel® Data Center GPU Max Series).

  • New supported workloads for Intel® Xeon® Scalable Processors:

    • PyTorch: Vision Transformer, GPT-J 6B, Llama2 (7B and 13B), ChatGLMv3 6B, LCM, YOLOv7
    • TensorFlow: Bert Large Hugging Face, GraphSAGE
  • New supported workloads for Intel® Data Center GPU platforms:

    • PyTorch: FBNet, IFRNet, RIFE
  • Added support for SRF: Updated PyTorch workload scripts for the best known configurations on CPU platforms (Intel® Xeon® Scalable Processors ).

  • Added support for TensorFlow+ XLA.

  • Fixed all CVEs and security issues.

Supported Configurations

Intel® AI Reference Models v3.2 is validated on the following environment:

  • Ubuntu 22.04 LTS
  • Ubuntu 20.04 LTS
  • Windows 11
  • Windows Subsystem for Linux 2 (WSL2)
  • Python 3.9, 3.10, 3.11

Intel® AI Reference Models v3.1.1

07 Mar 23:00
Compare
Choose a tag to compare

Supported Frameworks

New features

  • Added new workloads scripts for the best known configurations on GPU platforms (Intel® Data Center GPU Flex Series and Intel® Data Center GPU Max Series).

  • New supported workloads and precisions for Intel® Data Center GPU platforms:

    • Swin-Transformer, FastPitch, UNet++ and RNNT for Inference for Intel® Extension for PyTorch.
    • Updated YOLOv5, DLRM v2, and 3D-Unet for Inference for Intel® Extension for PyTorch.
  • Updated workload scripts for the best known configurations on CPU platforms (Intel® Xeon® Scalable Processors ).

Supported Configurations

Intel® AI Reference Models v3.1.1 is validated on the following environment:

  • Ubuntu 22.04 LTS
  • Ubuntu 20.04 LTS
  • Windows 11
  • Windows Subsystem for Linux 2 (WSL2)
  • Python 3.9, 3.10

Intel® AI Reference Models v3.1.0

27 Jan 01:14
Compare
Choose a tag to compare

Supported Frameworks

New features

  • Updated workloads scripts for the best known configurations on GPU platforms (Intel® Data Center GPU Flex Series, Intel® Data Center GPU Max Series and Intel® Arc™ A-Series Graphics).

  • Experimental support for Intel® Arc™ A-Series (Intel® Arc™ A770 Graphic card) GPUs on Windows Subsystem for Linux 2 with Ubuntu Linux installed and native Ubuntu Linux.

  • New supported workloads and precisions for Intel® Data Center GPU platforms:

    • Wide and Deep Large Model Inference and EfficientNet B4 for Intel® Extension for TensorFlow
    • DistilBert and DLRM v1 Inference for Intel® Extension for PyTorch
    • 3D-Unet for Intel® Extension for TensorFlow and Intel® Extension for PyTorch
  • Updated workload scripts for the best known configurations on CPU platforms (Intel® Xeon® Scalable Processors ).

  • Updated Transfer Learning Jupyter notebooks.

  • This release contains many bug and CVE fixes to the previous versions.

Supported Configurations

Intel® AI Reference Models v3.1.0 is validated on the following environment:

  • Ubuntu 22.04 LTS
  • Ubuntu 20.04 LTS
  • Windows 11
  • Windows Subsystem for Linux 2 (WSL2)
  • Python 3.9, 3.10

Intel® AI Reference Models v3.0.0

19 Oct 16:23
Compare
Choose a tag to compare

New features

  • Model Zoo for Intel® Architecture is rebranded as Intel® AI Reference Models to reflect it's purpose of showcasing to external audiences the internally achieved best performance configurations for critical workloads on Intel® Architecture.

  • Updates Intel® Data Center GPU Max Series 1550 x4 OAM workloads scripts for the best known configurations.

  • Introduces the initial version of Jupyter notebook interface for Intel® AI Reference Models.

  • Distributed training is supported for the following PyTorch models with different precisions: ResNet50, SSD-ResNet34, DLRM, MaskRCNN, RNNT and BERT Large.

  • Updated Transfer Learning Jupyter notebooks.

  • New supported workloads:

    • MemRec DLRM FP32 inference: introducing Memory Efficient Recommendation System using Alternative Representation (MemRec). MemRec is a technology for alternative representation of embedding tables. It uses bloom filters and hashing techniques to encode embedding tables in a much smaller memory footprint optimized to make use of hierarchical cache architecture of Intel Xeon platforms. MemRec encodes DLRM embedding tables into two cache-friendly embedding tables to maximize predictive performance and increase recommendation accuracy.
    • DLRM v2 training and inference with different precisions for GPU and CPU platforms.
  • This release contains many bug and CVE fixes to the previous versions.

Supported Configurations

Intel® AI Reference Models v3.0.0 is validated on the following environment:

  • Ubuntu 22.04 LTS
  • Ubuntu 20.04 LTS
  • Windows 11
  • Windows Subsystem for Linux 2 (WSL2)
  • Python 3.9, 3.10

Model Zoo for Intel® Architecture v2.12.1

15 Sep 22:57
Compare
Choose a tag to compare

New features

  • Updated Intel® Data Center GPU Flex 140 and 170 workloads scripts for the best known configurations.
  • New supported workloads for Intel® Data Center GPU Flex:
    • EfficientNet B0 and B3 for Intel® Extension for TensorFlow
    • MaskRCNN for Intel® Extension for TensorFlow
    • Stable Diffusion for Intel® Extension for TensorFlow and Intel® Extension for PyTorch
    • YOLO v5 for Intel® Extension for PyTorch

Deprecation Notice:

The next release of Model Zoo for Intel® Architecture will be rebranded as Intel® AI Reference Models to reflect it's purpose of showcasing to external audiences the internally achieved best performance configurations for critical workloads on Intel® Architecture.

Bug fixes:

Supported Configurations

Intel Model Zoo v2.12.1 is validated on the following environment:

  • Ubuntu 22.04 LTS
  • Ubuntu 20.04 LTS
  • Windows 11
  • Windows Subsystem for Linux 2 (WSL2)
  • Python 3.9, 3.10

Model Zoo for Intel® Architecture v2.11.1

21 Jul 17:19
Compare
Choose a tag to compare

New features

  • Updated Intel® Data Center GPU Flex and Max Series workloads scripts for the best known configurations.
  • Single dockerfile support for both Intel® Data Center GPU Flex 170 and Intel® Data Center GPU Flex 140.
  • Updated the name and license for the Cloud Data Connector.
  • Added comparative samples between cloud providers SDK and Cloud Data Connector.
  • Added the source code for the initial releases of the dataset-librarian python library and dataset-librarian Anaconda package.
  • Bump transformers from 4.25.1 to 4.30.0 to fix CVE.

Bug fixes:

Supported Configurations

Intel Model Zoo v2.11.1 is validated on the following environment:

  • Ubuntu 22.04 LTS
  • Ubuntu 20.04 LTS
  • Windows 11
  • Windows Subsystem for Linux 2 (WSL2)
  • Python 3.9, 3.10

Model Zoo for Intel® Architecture v2.11.0

28 Apr 23:16
Compare
Choose a tag to compare

Supported Frameworks

New models

  • New precisions FP16 and BFloat16 for different workloads

New features

Bug fixes:

Supported Configurations

Intel Model Zoo v2.11.0 is validated on the following environment:

  • Ubuntu 22.04 LTS
  • Ubuntu 20.04 LTS
  • Windows 11
  • Windows Subsystem for Linux 2 (WSL2)
  • Python 3.8, 3.9