Quantization Workflow Prerequisites - MATLAB & Simulink (original) (raw)

Main Content

This page describes the products required to quantize, simulate, and deploy deep learning networks using Deep Learning Toolbox Model Compression Library. The prerequisites required depend on your selections at each stage of the quantization workflow.

Prerequisites for All Quantization Workflows

The following requirements apply to all stages of the quantization workflow.

Deep Learning Toolbox™
Deep Learning Toolbox Model Compression Library

Supported Networks and Layers

The target hardware associated with each execution environment determines the supported networks and layers for that environment. For a list of supported layers, seeSupported Layers for Quantization.

Prerequisites for Calibration

The prerequisites for calibration depend on your selection of calibration environment.

Calibrate on host GPU —
- Parallel Computing Toolbox™
- GPU Coder™ Interface for Deep Learning
- CUDA® enabled NVIDIA® GPU with compute capability 3.5 or higher.
- Supported C/C++ Compiler
  * On Windows®, the MinGW C/C++ compiler is not supported. Use Microsoft Visual C++ 2022,Microsoft Visual C++ 2019, orMicrosoft Visual C++ 2017.
  * On Linux®, use a GCC C/C++ compiler.
  For a list of supported compilers, see Supported and Compatible Compilers.
Calibrate on host CPU —
- MATLAB® Coder™ Interface for Deep Learning or GPU Coder Interface for Deep Learning

Prerequisites for Quantization

The following are required to quantize your network for simulation in MATLAB using the quantize function or the option in the Deep Network Quantizer app.

When you use the MATLAB Execution Environment, a Fixed-Point Designer™ license is required for quantization.
When you use the CPU Execution Environment, the quantization step is not supported.
When you use any other Execution Environment for quantization, no additional prerequisites are required.

Prerequisites for Validation

The following are required to validate your quantized network for deployment using thevalidate function or the Validate button in the Deep Network Quantizer app.

Execution Environment	Prerequisites for Validation
GPU	Parallel Computing ToolboxGPU Coder Interface for Deep LearningCUDA enabled NVIDIA GPU with compute capability 6.1, 6.3 or higher. Setting Up the Prerequisite Products (GPU Coder)
FPGA	Deep Learning HDL Toolbox™Deep Learning HDL Toolbox Support Package for AMD FPGA and SoC DevicesDeep Learning HDL Toolbox Support Package for Intel® FPGA and SoC Deviceshdlsetuptoolpath (HDL Coder)
CPU	MATLAB Coder Interface for Deep LearningMATLAB CoderEmbedded Coder®ARM Compute Library. For more information, see Prerequisites for Deep Learning with MATLAB Coder (MATLAB Coder)
MATLAB	Fixed-Point Designer

For CPU and GPU deployment, the software generates code for a deep neural network by quantizing the weights, biases, and activations of the convolution layers to 8-bit scaled integer data types. The quantization is performed by providing the calibration result file produced by the calibrate function to the codegen (MATLAB Coder) command.

Code generation does not support quantized deep neural networks produced by thequantize function.

Validation on target hardware for CPU, FPGA, and GPU execution environments is not supported in MATLAB Online™. For FPGA and GPU execution environments, validation can be performed through emulation on the MATLAB Online host. GPU validation can also be performed if GPU support has been added to your MATLAB Online Server™ cluster. For more information on GPU support for MATLAB Online, see Configure GPU Support in MATLAB Online Server (MATLAB Online Server).