TFLiteModel

TensorFlow Lite model

Since R2022a

Description

Add-On Required: This feature requires the Deep Learning Toolbox Interface for LiteRT Library add-on.

A TFLiteModel object enables support for simulation and code generation for deep learning inference by using TensorFlow™ Lite models

Use a TFLiteModel object with the predict function in your MATLAB^® code to perform inference in MATLAB execution, code generation, or MATLAB Function block in Simulink^® models. For more information, see Prerequisites for Deep Learning with TensorFlow Lite Models.

Creation

To create a TFLiteModel object from a pretrained TensorFlow Lite model file, use the loadTFLiteModel function.

Properties

ModelName — Name of the TensorFlow Lite model file

character vector

Name of the TensorFlow Lite model file, specified as a character vector.

NumInputs — Number of inputs of the TensorFlow Lite model

numeric scalar

Number of inputs the TensorFlow Lite model accepts, specified as an integer-valued numeric scalar.

NumOutputs — Number of outputs of the TensorFlow Lite model

numeric scalar

Number of outputs the TensorFlow Lite model produces, specified as an integer-valued numeric scalar.

InputSize — Size of inputs of the TensorFlow model

cell array containing numeric arrays

Size of inputs of the TensorFlow model, specified as a cell array containing numeric arrays.

OutputSize — Size of outputs of the TensorFlow model

cell array containing numeric arrays

Size of outputs of the TensorFlow model, specified as a cell array containing numeric arrays.

NumThreads — Number of computational threads used for inference computation

Read-only: numeric scalar

This property is read-only.

Number of computational threads used for running inference with the TensorFlow Lite model, specified as an integer-valued numeric scalar.

To set the number of computational threads used for inference computation, specify the value when loading the TFLite model by using the NumThreads argument of the loadTFLiteModel function.

Mean — Mean for input data normalization

127.5 (default) | double scalar

Mean value to which the input data is normalized, specified as a double scalar. If the input data is not normalized, you must set this property to 0. Otherwise, set this property based on how the input data is normalized.

StandardDeviation — Standard deviation for input data normalization

127.5 (default) | double scalar

Standard deviation to which the input data is normalized, specified as a double scalar. If the input data is not normalized, you must set this property to 1. Otherwise, set this property based on how the input data is normalized.

Object Functions

predict

Compute deep learning network output for inference by using a TensorFlow Lite model

Examples

Perform Inference with TensorFlow Lite Model

Suppose that your current working directory contains a TensorFlow Lite Model named mobilenet_v1_0.5_224.tflite.

Load the model by using the loadTFLite function. Inspect the object this function creates.

net = loadTFLiteModel('mobilenet_v1_0.5_224.tflite');
disp(net)

  TFLiteModel with properties:
            ModelName: 'mobilenet_v1_0.5_224.tflite'
            NumInputs: 1
           NumOutputs: 1
            InputSize: {[224 224 3]}
           OutputSize: {[1001 1]}
           NumThreads: 8
                 Mean: 127.5000
    StandardDeviation: 127.5000

Create a MATLAB function that can perform inference using the object net. This function loads the Mobilenet-V1 model into a persistent network object. Then the function performs prediction by passing the network object to the predict function. Subsequent calls to this function reuse this the persistent object.

function out = tflite_predict(in)
persistent net;
if isempty(net)
    net = loadTFLiteModel('mobilenet_v1_0.5_224.tflite');
end
out = predict(net,in);
end

For an example that shows how to generate code for this function and deploy on Raspberry Pi^® hardware, see Generate Code for TensorFlow Lite (TFLite) Model and Deploy on Raspberry Pi.

Extended Capabilities

C/C++ Code Generation

Generate C and C++ code using MATLAB® Coder™.

Usage notes and limitations:

• Only C++ code generation is supported.

• You cannot modify the NumThreads property of a TFLiteModel object after it is created. To optimize performance, specify the number of computational threads when you load the TFLite model by using the loadTFLiteModel function.

• If code generation target is the host computer, the default value of the NumThreads property is equal to the value returned by the maxNumCompThreads function.

• If code generation target is a non-host hardware board, the default value of the NumThreads property is 4.

Version History

Introduced in R2022a

R2026a: Property NumThreads is read-only

You cannot modify the NumThreads property of a TFLiteModel object.

To set the number of computational threads used for inference computation, you must specify the value when loading the TFLite model by using the NumThreads argument of the loadTFLiteModel function.