[go: nahoru, domu]

Segmentation Platform

Introduction

The segmentation platform is a platform that uses intelligence and machine learning to guide developers for building purpose-built user experience for specific segments of users.

Segmentation Platform is a layered component (https://sites.google.com/a/chromium.org/dev/developers/design-documents/layered-components-design) to enable it to be easily used on all platforms.

Code structure

components/segmentation_platform/public Public interfaces and data structure.

components/segmentation_platform/internal Internal implementations.

chrome/browser/segmentation_platform Includes factories to instantiate the service.

SegmentationPlatformService - Public interface for segmentation platform service.

Test models

components/test/data/segmentation_platform contains ML models used for testing.

  • adder.tflite: Takes two floats as input in a single tensor. Outputs a single tensor with a single element which is the sum of the two floats given as input.

Testing

To run all the relevant C++ unit tests, you can run the components_unittests target and give the segmentation_platform filter file as an argument:

./out/Default/components_unittests --test-launcher-filter-file=components/segmentation_platform/components_unittests.filter

To update the list of tests, you can run the following command:

components/segmentation_platform/tools/testing/launcher_filter_file.py