Version History/Revision History

This release marks the second set of updates and new features in the Edge Developer Toolbox since the public launch at Mobile World Congress 2024.

Introduction

Intel continues to build the Edge Developer Toolbox as a seamless one-stop workspace to help developers create innovative Edge AI solutions and get hands-on experience building with the Intel® Tiber™ Edge Platform.

Features in This Release

• Microservices Integration - For microservices available on Edge Services Catalog, a “Try on Edge Developer Toolbox” option is now available. Users can now gain practical experience with microservices using these options through JupyterLab* notebooks. This hands-on approach allows users to explore and understand the capabilities of various microservices. The following microservices are available in the catalog, which is integrated with the Edge Developer Toolbox:

  • Edge Video Analytics Microservice

  • Data Store Microservice

  • Intel® Edge Data Collection

  • Deep Learning Workbench

    • Hugging Face Gen AI Models: Users can now import and benchmark over 30 models from Hugging Face on Intel platforms using the Deep Learning Workbench.

    • Post-Training INT8 Quantization: The Deep Learning Workbench now supports INT8 quantization, allowing users to quantize any PyTorch models with FP32, FP16 precision to INT8 for faster inference. This results in a lower memory footprint and improved performance. Users can benchmark these INT8 quantized models on Intel platforms, comparing performance gains before and after quantization using telemetry data visualization.

    For more information, refer to the the OpenVINO™ toolkit Model Optimization Guide: https://docs.openvino.ai/2024/openvino-workflow/model-optimization-guide/quantizing-models-post-training/basic-quantization-flow.html