class TensorflowLite::Client

• TensorflowLite::Client
• Reference
• Object

Overview

provides a simplified way to load and manipulate the tensorflow interpreter

the indexable module provides simplified access to the input tensors

Included Modules

• Indexable(TensorflowLite::Tensor)

Defined in:

tensorflow_lite/client.cr

Constructors

• .new(model : URI | Bytes | Path | Model | String, delegate : Delegate | Nil = nil, threads : Int | Nil = nil, labels : URI | Array(String) | Nil = nil)

  Configures the tensorflow interpreter with the options provided

• .new(model : URI | Bytes | Path | Model | String, delegate : Delegate | Nil = nil, threads : Int | Nil = nil, labels : URI | Array(String) | Nil = nil, &on_error : String -> Nil)

  Configures the tensorflow interpreter with the options provided

Instance Method Summary

• #input_tensor(*args, **options)

  input tensor details for manipulation and loading of input data

• #input_tensor(*args, **options, &)

  input tensor details for manipulation and loading of input data

• #input_tensor_count(*args, **options)

  input tensor details for manipulation and loading of input data

• #input_tensor_count(*args, **options, &)

  input tensor details for manipulation and loading of input data

• #interpreter : Interpreter
• #invoke(*args, **options)

  run the model, processing the input tensors and updating the output tensors

• #invoke(*args, **options, &)

  run the model, processing the input tensors and updating the output tensors

• #invoke!(*args, **options)

  run the model, processing the input tensors and updating the output tensors

• #invoke!(*args, **options, &)

  run the model, processing the input tensors and updating the output tensors

• #labels

  attempt to extract any labels in the model

• #labels_fetched : Bool
• #model : Model
• #model_path : Path | Nil
• #num_threads(*args, **options)

  This controls the number of CPU threads that the interpreter will use for its computations.

• #num_threads(*args, **options, &)

  This controls the number of CPU threads that the interpreter will use for its computations.

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

  provide a callback to receive any error messages

• #options : InterpreterOptions
• #output(index : Int = 0)

  returns the output tensor at the provided index

• #output_tensor(*args, **options)

  output tensors details, used to obtain the results of an invokation

• #output_tensor(*args, **options, &)

  output tensors details, used to obtain the results of an invokation

• #output_tensor_count(*args, **options)

  output tensors details, used to obtain the results of an invokation

• #output_tensor_count(*args, **options, &)

  output tensors details, used to obtain the results of an invokation

• #outputs

  returns an array of output tensors

• #size

  the number of input tensors

Constructor Detail

Instance Method Detail