class TensorflowLite::InterpreterOptions

Overview

Interpreter options provide a way to configure various aspects of the TensorFlow Lite runtime

Defined in:

tensorflow_lite/interpreter_options.cr

Constructors

.new

Instance Method Summary

#add_delegate(delegate : Delegate)
Delegates are mechanisms that allow the interpreter to offload some or all of the model execution to hardware accelerators, like the GPU, DSP, or specialized Neural Processing Units (NPUs).
#num_threads(count : Int)
This controls the number of CPU threads that the interpreter will use for its computations.
#on_error(&callback : String -> Nil)
define a callback to receive any error messages

Constructor Detail

def self.new #

[View source]

Instance Method Detail

def add_delegate(delegate : Delegate) #

Delegates are mechanisms that allow the interpreter to offload some or all of the model execution to hardware accelerators, like the GPU, DSP, or specialized Neural Processing Units (NPUs).

[View source]

def num_threads(count : Int) #

This controls the number of CPU threads that the interpreter will use for its computations. This can be useful for improving performance on devices with multiple CPU cores.

[View source]

def on_error(&callback : String -> Nil) #

define a callback to receive any error messages

[View source]