openvino_genai.Text2SpeechPipeline#

class openvino_genai.Text2SpeechPipeline#

Bases: pybind11_object

Text-to-speech pipeline

__init__(self: openvino_genai.py_openvino_genai.Text2SpeechPipeline, models_path: os.PathLike | str | bytes, device: str, **kwargs) → None#: Text2SpeechPipeline class constructor. models_path (os.PathLike): Path to the model file. device (str): Device to run the model on (e.g., CPU, GPU).

Methods

`__delattr__`(name, /)	Implement delattr(self, name).
`__dir__`()	Default dir() implementation.
`__eq__`(value, /)	Return self==value.
`__format__`(format_spec, /)	Default object formatter.
`__ge__`(value, /)	Return self>=value.
`__getattribute__`(name, /)	Return getattr(self, name).
`__getstate__`()	Helper for pickle.
`__gt__`(value, /)	Return self>value.
`__hash__`()	Return hash(self).
`__init__`(self, models_path, device, **kwargs)	Text2SpeechPipeline class constructor.
`__init_subclass__`	This method is called when a class is subclassed.
`__le__`(value, /)	Return self<=value.
`__lt__`(value, /)	Return self<value.
`__ne__`(value, /)	Return self!=value.
`__new__`(**kwargs)
`__reduce__`()	Helper for pickle.
`__reduce_ex__`(protocol, /)	Helper for pickle.
`__repr__`()	Return repr(self).
`__setattr__`(name, value, /)	Implement setattr(self, name, value).
`__sizeof__`()	Size of object in memory, in bytes.
`__str__`()	Return str(self).
`__subclasshook__`	Abstract classes can override this to customize issubclass().
`_pybind11_conduit_v1_`
`generate`(args, *kwargs)	Overloaded function.
`get_generation_config`(self)
`get_speaker_embedding_shape`(self)	Get the expected speaker embedding shape for the loaded model.
`set_generation_config`(self, config)

Attributes

__annotations__

__annotations__ = {}#

__class__#: alias of pybind11_type

__delattr__(name, /)#: Implement delattr(self, name).

__dir__()#: Default dir() implementation.

__eq__(value, /)#: Return self==value.

__format__(format_spec, /)#

Default object formatter.

Return str(self) if format_spec is empty. Raise TypeError otherwise.

__ge__(value, /)#: Return self>=value.

__getattribute__(name, /)#: Return getattr(self, name).

__getstate__()#: Helper for pickle.

__gt__(value, /)#: Return self>value.

__hash__()#: Return hash(self).

__init__(self: openvino_genai.py_openvino_genai.Text2SpeechPipeline, models_path: os.PathLike | str | bytes, device: str, **kwargs) → None#: Text2SpeechPipeline class constructor. models_path (os.PathLike): Path to the model file. device (str): Device to run the model on (e.g., CPU, GPU).

__init_subclass__()#

This method is called when a class is subclassed.

The default implementation does nothing. It may be overridden to extend subclasses.

__le__(value, /)#: Return self<=value.

__lt__(value, /)#: Return self<value.

__ne__(value, /)#: Return self!=value.

__new__(**kwargs)#

__reduce__()#: Helper for pickle.

__reduce_ex__(protocol, /)#: Helper for pickle.

__repr__()#: Return repr(self).

__setattr__(name, value, /)#: Implement setattr(self, name, value).

__sizeof__()#: Size of object in memory, in bytes.

__str__()#: Return str(self).

__subclasshook__()#

Abstract classes can override this to customize issubclass().

This is invoked early on by abc.ABCMeta.__subclasscheck__(). It should return True, False or NotImplemented. If it returns NotImplemented, the normal algorithm is used. Otherwise, it overrides the normal algorithm (and the outcome is cached).

_pybind11_conduit_v1_()#

generate(*args, **kwargs)#

Overloaded function.

generate(self: openvino_genai.py_openvino_genai.Text2SpeechPipeline, text: str, speaker_embedding: object = None, **kwargs) -> openvino_genai.py_openvino_genai.Text2SpeechDecodedResults

Generates speeches based on input texts

param text_or_texts:

input text(s) for which to generate speech

type text_or_texts:

str or list[str]

:param speaker_embedding optional speaker embedding tensor representing the unique characteristics of a speaker’s
voice. If not provided for SpeechT5 TSS model, the 7306-th vector from the validation set of the Matthijs/cmu-arctic-xvectors dataset is used by default. Kokoro backend requires callers to prepare this tensor externally and pass it explicitly.

type speaker_embedding:

openvino.Tensor or None

param properties:

speech generation parameters specified as properties

type properties:

dict

returns:

raw audios of the input texts spoken in the specified speaker’s voice; sample rate is provided via Text2SpeechDecodedResults.output_sample_rate

rtype:

Text2SpeechDecodedResults

SpeechGenerationConfig

Speech-generation specific parameters: :param minlenratio: minimum ratio of output length to input text length; prevents output that’s too short. :type minlenratio: float

param maxlenratio:

maximum ratio of output length to input text length; prevents excessively long outputs.

type maxlenratio:

float

param threshold:

probability threshold for stopping decoding; when output probability exceeds above this, generation will stop.

type threshold:

float

Kokoro-specific parameters: :param speed: speech speed multiplier. :type speed: float

param language:

language code for Kokoro G2P (for example, “en-us” or “en-gb”).

type language:

str

param max_phoneme_length:

maximum phoneme chunk length for Kokoro preprocessing.

type max_phoneme_length:

int

param phonemize_fallback_model_dir:

Optional OpenVINO fallback phonemizer model directory. This applies only to fallback during phonemize / G2P (graphemes to phonemes), before acoustic model inference. If set, this OpenVINO G2P fallback is used. If unset (None), espeak-ng G2P fallback is used. For kwargs-based APIs (SpeechGenerationConfig(**kwargs), update_generation_config(**kwargs), and pipeline kwargs), omit this key instead of passing None because kwargs-to-AnyMap conversion rejects None values.

type phonemize_fallback_model_dir:

str | None
generate(self: openvino_genai.py_openvino_genai.Text2SpeechPipeline, texts: collections.abc.Sequence[str], speaker_embedding: object = None, **kwargs) -> openvino_genai.py_openvino_genai.Text2SpeechDecodedResults

Generates speeches based on input texts

param text_or_texts:

input text(s) for which to generate speech

type text_or_texts:

str or list[str]

:param speaker_embedding optional speaker embedding tensor representing the unique characteristics of a speaker’s
voice. If not provided for SpeechT5 TSS model, the 7306-th vector from the validation set of the Matthijs/cmu-arctic-xvectors dataset is used by default. Kokoro backend requires callers to prepare this tensor externally and pass it explicitly.

type speaker_embedding:

openvino.Tensor or None

param properties:

speech generation parameters specified as properties

type properties:

dict

returns:

raw audios of the input texts spoken in the specified speaker’s voice; sample rate is provided via Text2SpeechDecodedResults.output_sample_rate

rtype:

Text2SpeechDecodedResults

SpeechGenerationConfig

Speech-generation specific parameters: :param minlenratio: minimum ratio of output length to input text length; prevents output that’s too short. :type minlenratio: float

param maxlenratio:

maximum ratio of output length to input text length; prevents excessively long outputs.

type maxlenratio:

float

param threshold:

probability threshold for stopping decoding; when output probability exceeds above this, generation will stop.

type threshold:

float

Kokoro-specific parameters: :param speed: speech speed multiplier. :type speed: float

param language:

language code for Kokoro G2P (for example, “en-us” or “en-gb”).

type language:

str

param max_phoneme_length:

maximum phoneme chunk length for Kokoro preprocessing.

type max_phoneme_length:

int

param phonemize_fallback_model_dir:

Optional OpenVINO fallback phonemizer model directory. This applies only to fallback during phonemize / G2P (graphemes to phonemes), before acoustic model inference. If set, this OpenVINO G2P fallback is used. If unset (None), espeak-ng G2P fallback is used. For kwargs-based APIs (SpeechGenerationConfig(**kwargs), update_generation_config(**kwargs), and pipeline kwargs), omit this key instead of passing None because kwargs-to-AnyMap conversion rejects None values.

type phonemize_fallback_model_dir:

str | None

get_generation_config(self: openvino_genai.py_openvino_genai.Text2SpeechPipeline) → openvino_genai.py_openvino_genai.SpeechGenerationConfig#

get_speaker_embedding_shape(self: openvino_genai.py_openvino_genai.Text2SpeechPipeline) → openvino._pyopenvino.Shape#: Get the expected speaker embedding shape for the loaded model. SpeechT5: Shape{1, 512}. Kokoro: Shape{510, 1, 256}

set_generation_config(self: openvino_genai.py_openvino_genai.Text2SpeechPipeline, config: openvino_genai.py_openvino_genai.SpeechGenerationConfig) → None#