Documentation

Scope Fusion Pattern Classification

A collection of built-in scope fusion patterns is provided to developers, as introduced in TensorFlow Parser Scope Fusion Patterns. This guide gives step-by-step instructions for customizing scope fusion patterns by using the available APIs.

Scope fusion patterns, built-in or custom, are classified into the following two types:
  • General: common scope fusion patterns applicable to all networks. They are enabled by default and cannot be manually disabled.
  • Non-general: fusion patterns applicable to specific networks. By default, they are disabled. During ATC model conversion or TensorFlow graph execution, you can enable the non-general fusion patterns as required.
Parent topic: Basic Concepts