tf.lite.TFLiteConverter

Converts a TensorFlow model into TensorFlow Lite model.

tf.lite.TFLiteConverter(
    funcs, trackable_obj=None
)

Example usage:

# Converting a SavedModel to a TensorFlow Lite model.
converter = tf.lite.TFLiteConverter.from_saved_model(saved_model_dir)
tflite_model = converter.convert()

# Converting a tf.Keras model to a TensorFlow Lite model.
converter = tf.lite.TFLiteConverter.from_keras_model(model)
tflite_model = converter.convert()

# Converting ConcreteFunctions to a TensorFlow Lite model.
converter = tf.lite.TFLiteConverter.from_concrete_functions([func], model)
tflite_model = converter.convert()

# Converting a Jax model to a TensorFlow Lite model.
converter = tf.lite.TFLiteConverter.experimental_from_jax(
    [func], [[ ('input1', input1), ('input2', input2)]])
tflite_model = converter.convert()

funcs List of TensorFlow ConcreteFunctions. The list should not contain duplicate elements.
trackable_obj tf.AutoTrackable object associated with funcs. A reference to this object needs to be maintained so that Variables do not get garbage collected since functions have a weak reference to Variables. This is only required when the tf.AutoTrackable object is not maintained by the user (e.g. from_saved_model).

optimizations Experimental flag, subject to change. Set of optimizations to apply. e.g {tf.lite.Optimize.DEFAULT}. (default None, must be None or a set of values of type tf.lite.Optimize)
representative_dataset A generator function used for integer quantization where each generated sample has the same order, type and shape as the inputs to the model. Usually, this is a small subset of a few hundred samples randomly chosen, in no particular order, from the training or evaluation dataset. This is an optional attribute, but required for full integer quantization, i.e, if tf.int8 is the only supported type in target_spec.supported_types. Refer to tf.lite.RepresentativeDataset. (default None)
target_spec Experimental flag, subject to change. Specifications of target device, including supported ops set, supported types and a set of user's defined TensorFlow operators required in the TensorFlow Lite runtime. Refer to tf.lite.TargetSpec.
inference_input_type Data type of the input layer. Note that integer types (tf.int8 and tf.uint8) are currently only supported for post training integer quantization and quantization aware training. (default tf.float32, must be in {tf.float32, tf.int8, tf.uint8})
inference_output_type Data type of the output layer. Note that integer types (tf.int8 and tf.uint8) are currently only supported for post training integer quantization and quantization aware training. (default tf.float32, must be in {tf.float32, tf.int8, tf.uint8})
allow_custom_ops Boolean indicating whether to allow custom operations. When False, any unknown operation is an error. When True, custom ops are created for any op that is unknown. The developer needs to provide these to the TensorFlow Lite runtime with a custom resolver. (default False)
exclude_conversion_metadata Whether not to embed the conversion metadata into the converted model. (default False)
experimental_new_converter Experimental flag, subject to change. Enables MLIR-based conversion. (default True)
experimental_new_quantizer Experimental flag, subject to change. Enables MLIR-based quantization conversion instead of Flatbuffer-based conversion. (default True)
experimental_enable_resource_variables Experimental flag, subject to change. Enables resource variables to be converted by this converter. This is only allowed if the from_saved_model interface is used. (default True)

Methods

convert

View source

convert()

Converts a TensorFlow GraphDef based on instance variables.

Returns
The converted data in serialized format.

Raises
ValueError No concrete functions is specified. Multiple concrete functions are specified. Input shape is not specified. Invalid quantization parameters.

experimental_from_jax

View source

@classmethod
experimental_from_jax(
    serving_funcs, inputs
)

Creates a TFLiteConverter object from a Jax model with its inputs.

Args
serving_funcs A array of Jax functions with all the weights applied already.
inputs A array of Jax input placeholders tuples list, e.g., jnp.zeros(INPUT_SHAPE). Each tuple list should correspond with the serving function.

Returns
TFLiteConverter object.

from_concrete_functions

View source

@classmethod
from_concrete_functions(
    funcs, trackable_obj=None
)

Creates a TFLiteConverter object from ConcreteFunctions.

Args
funcs List of TensorFlow ConcreteFunctions. The list should not contain duplicate elements. Currently converter can only convert a single ConcreteFunction. Converting multiple functions is under development.
trackable_obj An AutoTrackable object (typically tf.module) associated with funcs. A reference to this object needs to be maintained so that Variables do not get garbage collected since functions have a weak reference to Variables.

Returns
TFLiteConverter object.

Raises
Invalid input type.

from_keras_model

View source

@classmethod
from_keras_model(
    model
)

Creates a TFLiteConverter object from a Keras model.

Args
model tf.Keras.Model

Returns
TFLiteConverter object.

from_saved_model

View source

@classmethod
from_saved_model(
    saved_model_dir, signature_keys=None, tags=None
)

Creates a TFLiteConverter object from a SavedModel directory.

Args
saved_model_dir SavedModel directory to convert.
signature_keys List of keys identifying SignatureDef containing inputs and outputs. Elements should not be duplicated. By default the signatures attribute of the MetaGraphdef is used. (default saved_model.signatures)
tags Set of tags identifying the MetaGraphDef within the SavedModel to analyze. All tags in the tag set must be present. (default {tf.saved_model.SERVING} or {'serve'})

Returns
TFLiteConverter object.

Raises
Invalid signature keys.