Welcome to ONE’s documentation!

• Overview
  • Background
  • Roadmap
    • Progress
    • Goal
    • Deliverables
    • Milestones
    • Workgroups (WGs)
  • Overall Architecture
    • Compiler
    • Package
    • Runtime
  • Supported Operations
    • circle additional operators
  • Workgroup
    • Runtime WG1
    • Runtime WG2
    • Runtime WG3
    • Compiler Frontend WG
    • Compiler Backend WG
    • How to Join?
• How To
  • How to Build Compiler
    • Build Requires
    • Build for Ubuntu
      • Build for debug and release separately
    • Build for Windows
    • Cross build for Ubuntu/ARM32 (experimental)
      • Prepare root file system
      • Clean existing external source for patches
      • Build
      • Test
    • Known issues
  • How to Build Package
    • Overview
    • Import model
      • Example for TensorFlow
      • Example for TensorFlow lite
      • Example for TensorFlow Model Including BCQ Information
    • Optimize circle model
    • Quantize circle model
    • Pack circle model
  • How to Build Runtime
    • Build requirements
    • Get ONE source code
    • Install python packages
    • Build from source code, for Ubuntu
      • Run test
    • Build for Tizen
    • Build using docker image
    • How to cross-build runtime for different architecture
  • How to Build Runtime with GBS for Tizen/RPi4
    • Setting up build environment
    • Build ONERT
    • Prepare Tizen on Raspberry Pi 4
    • sdb connect to Tizen/RPi4
  • How to Build Using Prebuilt Docker Image
  • How to Cross-build Runtime for AARCH64
    • Prepare AARCH64 Ubuntu RootFS
      • Prepare RootFS at alternative folder
      • Using proxy
    • Cross build for AARCH64
      • Run test
  • How to Cross-build Runtime for Android
    • Prepare Android NDK
      • Install Android NDK
      • Host Environment Requirements
    • Build
      • Build and install the runtime
  • How to Cross-build Runtime for ARM
    • Prepare ARM Ubuntu RootFS
      • Prepare RootFS at alternative folder
      • Using proxy
    • Install ARM Cross Toolchain
    • Build and install ARM Compute Library
    • Install numpy
    • Cross build for ARM by using Makefile.template
      • Run test
  • How to Run Package
  • How to Make an Application with Runtime
  • How to Use API
  • How to Use NNAPI Binding
  • How to Use NNFW API
    • Prepare nnpackage
      • Convert tensorflow pb file to nnpackage
      • Convert tflite file to nnpackage
    • Build app with NNFW API
    • Run Inference with app on the target devices
  • How to Use Specific Backend during Inference
    • Using NNFW API
      • nnfw_set_available_backends
    • Using Environment Variable
      • 1. BACKENDS
      • 2. OP_BACKEND_[OP_TYPE]
      • 3. OP_BACKEND_MAP
  • How to Contribute
    • General contribution guidelines
    • How to create a Pull Request
  • How to Remote Debugging with Visual Studio Code
    • Setup build host
      • Install gdb-multiarch
      • Configure VS code
    • Setup target device
      • Install gdbserver and debugging symbols
    • Run remote debugging
    • Optional: Setup rootfs on build host
    • Troubleshooting
      • Unable to open ‘unordered_map.h’
      • Skip STL files
  • How to Add a New Operation
    • Compiler
    • Runtime
  • How To Introduce a New Operation Into Compiler
  • How To Introduce a New Operation Into Runtime
    • Index
    • Core
    • Frontend
      • Loaders
        • Base Loader
        • TFLite Loader
        • Circle Loader
      • NNAPI
    • Backend
      • ShapeFixer
        • acl_cl
        • acl_neon
        • cpu
      • KernelGenerator
        • acl_cl
        • acl_neon
        • cpu
      • ConstantInitializer (in some cases)
        • cpu
    • Samples (to be updated)
• Runtime
  • API
    • Runtime Layered Architecture
    • Frontend API
      • NN API
      • NNFW API
    • Backend API
  • Core
    • Modules
      • Module ir
      • Module compiler
        • 1. Lowering
        • 2. Tensor Registration
          • Q. What are the differences between ‘operand’ and ‘tensor’?
        • 3. Linearization (Linear Executor Only)
        • 4. Kernel Generation
        • 5. Create Executor
      • Module exec
      • Module backend
  • Controlflow Operations
    • Defining controlflow operations
    • Implementation
      • Graph representation
      • Execution
        • Kernel Implementation
  • Executors
    • Understanding models
    • Linear Executor
    • Dataflow Executor (experimental)
    • Parallel Executor (experimental)
  • Heterogeneous Execution
    • Graph Transformation
    • Automatic Scheduling
  • Backend API
    • How backends are loaded
    • C and C++ API
      • C API
      • C++ API
    • Provided Backend Implementations
    • builtin Backend
    • Bundle Backends
      • cpu
      • acl_neon
      • acl_cl
  • Compute
    • ARMComputeEx
    • cker
  • Supported Operations and backend
    • Raw-data format (float32, int32, boolean, etc)
    • Quantization format (uint8 asymmetric)
    • Quantization format (int8)
• Compiler
  • Frontend
    • circle2circle
    • tf2tfliteV2
      • Where does V2 come from?
      • Prerequisite
      • Example
      • Optional arguments
    • tflite2circle
      • Usage
  • Middleend
    • exo
      • How to add a new TFL node
    • locoex
    • logo
    • logo-core
  • Backend
  • IR
    • loco
    • Dialect Service
      • Revision
      • Motivation
      • Design
        • What is “service”?
        • How to declare a service
        • How to access a service
        • How to register a service
    • luci
    • luci-export
    • luci-import
    • luci-lang
    • luci-logex
    • luci-log
    • luci-pass
    • luci-service
    • oneco
  • Interpreter
    • locomotiv
    • Purpose
    • Sample code to use locomotiv library
    • How to support new loco node execution: recommended guide
      • Steps to support new loco node
        • Note on internal data layout rule
        • Notes on step 3
        • Recommendation on step 4 (test)
  • Libraries
    • adtidas
    • angkor
      • Purpose
      • How to use
      • Example
    • bino
    • cli
    • Background
    • How to use
    • cwrap
      • How to use
      • Example
    • fipe
    • hermes
    • hermes-std
    • kuma
      • What does “kuma” mean?
    • locop
    • mio-circle
    • mio-tf
    • mio-tflite
    • moco-log
    • morph
    • nest
      • References
    • nike
    • nnop
    • oops
    • pepper-assert
    • pepper-env
    • pepper-str
      • HOW TO USE
    • pepper-strcast
    • plier-tf
    • pp
    • Function (Feature)
    • How to use
    • safemain
    • v4tf
      • What is this?
      • Why do we need this?
    • dredd-rule-lib
      • Terms
      • Metrics supported
      • Related projects - dredd tests
      • Rule file
        • Naming convention of rule file
      • How do all these work?
      • Model repo and How to add a model as a target of a dredd-test.
    • tfinfo-v2
  • Tools
    • circle-inspect
      • Information to inspect
    • circle-verify
      • Usage
    • circlechef
      • What is circlechef?
    • circledump
      • What is this?
      • Design philosophy
      • To do
      • How to use
      • Dependency
    • i5diff
      • Yet Another Diff?
      • How to use
    • onnxkit
      • Purpose
      • How to use
      • Examples
      • Dependency
    • tfkit
      • What is tfkit?
      • Tutorial: How to use?
        • Examples
    • tfl-inspect
      • Information to inspect
        • –operators
        • –conv2d_weight
    • tfl-verify
      • Usage
    • tflchef
      • What is tflchef?
      • Tutorial: How to use?
    • tfldump
      • What is this?
      • Design philosophy
      • To do
      • How to use
      • Dependency
    • oneco-value-pbtxt-test
    • tf2tfliteV2-conversion-test
    • tflite2circle-conversion-test
• Common IR
  • Introduction to circle
  • What is Common IR
• Package
  • Introduction to Package
  • Design of Package
    • Manifest
    • Supported Models
    • User Defined Operation
• Platform
  • Ubuntu
  • Tizen
  • Android
• Devices
  • ODROID-XU3
  • ODROID-XU4
  • Raspberry Pi 3
• Test & Benchmarks
  • Scripts
  • Benchmarks
• Release
  • 1.0
    • Release Note 1.0.0
  • 1.1
    • Release Note 1.1.0
  • 1.2
    • Release Note 1.2.0
  • 1.3
    • Release Note 1.3.0
  • 1.4
    • Release Note 1.4.0
  • 1.5
    • Release Note 1.5.0
      • Feature Highlights
      • ONE Compiler
        • Compiler supports more operations
      • ONE Runtime
        • CPU backend supports more operations
  • 1.6
    • Release Note 1.6.0
      • Feature Highlights
      • ONE Compiler
        • Compiler supports more operations
      • ONE Runtime
        • CPU backend supports more operations
        • Support dynamically shaped tensors
        • Support Control Flow operations
        • API updates
  • 1.7
    • Feature Highlights
    • ONE Compiler
      • Compiler supports more operations
      • New command line interface for user interface consistancy
    • ONE Runtime
      • Runtime CPU backend supports more operations
      • Runtime CPU backend supports more quant8 operations
      • API changes
      • New optimization
  • 1.8
    • Release Note 1.8.0
      • Feature Highlights
      • ONE Compiler
        • New command line interface for user interface consistancy
        • Changes
      • ONE Runtime
        • Rename headers
        • Optimization
        • Support CPU backend operation
        • Support CPU backend quant8 operation
  • 1.9
    • Release Note 1.9.0
      • ONE Compiler
        • Compiler supports more operations
        • Changes
      • ONE Runtime
        • Runtime backend operation support
        • Optimization
        • Operation Implementation
        • Change
    • Release Note 1.9.1
      • ONE Compiler
        • Changes
  • 1.10
    • Release Note 1.10.0
      • ONE Compiler
        • Compiler supports more operations
        • Changes
      • ONE Runtime
        • Runtime backend operation support
        • Optimization
  • 1.11
    • Release Note 1.11.0
      • ONE Compiler
        • Compiler supports more operations
        • Changes
      • ONE Runtime
        • Runtime backend operation supports more operations and types
  • 1.12
    • Release Note 1.12.0
      • ONE Compiler
        • Compiler Frontend
      • ONE Runtime
        • Runtime backend operation supports more operations and types
        • nnpackage defines configuration file
  • 1.13
    • Release Note 1.13.0
      • ONE Compiler
        • Compiler Frontend
  • 1.14
    • Release Note 1.14.0
      • ONE Compiler
        • Compiler Frontend
  • 1.15
    • Release Note 1.15.0
      • ONE Compiler
        • Compiler Frontend
      • ONE Runtime
        • Runtime backend supports more operations and types
        • ARM Compute Library Update
  • 1.16
    • Release Note 1.16.0
      • ONE Compiler
        • Compiler Frontend
    • Release Note 1.16.1
      • ONE Compiler
        • Compiler Frontend

Indices and tables

• Index

• Module Index

• Search Page