Model Configs

Suggest Edits

FaceConfig

Configuration for the facial expression model.

@dataclass
class FaceConfig(ModelConfigBase["FaceConfig"])

Arguments:

  • fps_pred Optional[float] - Number of frames per second to process. Other frames will be omitted
    from the response.
    This configuration is only available for the batch API.
  • prob_threshold Optional[float] - Face detection probability threshold. Faces detected with a
    probability less than this threshold will be omitted from the response.
    This configuration is only available for the batch API.
  • identify_faces Optional[bool] - Whether to return identifiers for faces across frames.
    If true, unique identifiers will be assigned to face bounding boxes to differentiate different faces.
    If false, all faces will be tagged with an "unknown" ID.
  • min_face_size Optional[float] - Minimum bounding box side length in pixels to treat as a face.
    Faces detected with a bounding box side length in pixels less than this threshold will be
    omitted from the response.
    This configuration is only available for the batch API.
  • save_faces Optional[bool] - Whether to extract and save the detected faces to the artifacts
    directory included in the response.
    This configuration is only available for the batch API.
  • descriptions Optional[Dict[str, Any]] - Configuration for Descriptions predictions.
    Descriptions prediction can be enabled by setting "descriptions": {}.
    Currently, Descriptions prediction cannot be further configured with any parameters.
    If missing or null, no descriptions predictions will be generated.
  • facs Optional[Dict[str, Any]] - Configuration for FACS predictions.
    FACS prediction can be enabled by setting "facs": {}.
    Currently, FACS prediction cannot be further configured with any parameters.
    If missing or null, no facs predictions will be generated.

get_model_type

Get the configuration model type.

@classmethod
def get_model_type(cls) -> ModelType

Returns:

  • ModelType - Model type.

LanguageConfig

Configuration for the language emotion model.

@dataclass
class LanguageConfig(ModelConfigBase["LanguageConfig"])

Arguments:

  • granularity Optional[str] - The granularity at which to generate predictions.
    Accepted values are word, sentence, utterance, or conversational_turn.
    The default is utterance.
    utterance corresponds to a natural pause or break in conversation
    conversational_turn corresponds to a change in speaker.
    This configuration is available for the streaming API, but only with values word and sentence.
  • identify_speakers Optional[bool] - Whether to return identifiers for speakers over time.
    If true, unique identifiers will be assigned to spoken words to differentiate different speakers.
    If false, all speakers will be tagged with an "unknown" ID.
    This configuration is only available for the batch API.
  • sentiment Optional[Dict[str, Any]] - Configuration for Sentiment predictions.
    Sentiment prediction can be enabled by setting "sentiment": {}.
    Currently, Sentiment prediction cannot be further configured with any parameters.
    If missing or null, no sentiment predictions will be generated.
  • toxicity Optional[Dict[str, Any]] - Configuration for Toxicity predictions.
    Toxicity prediction can be enabled by setting "toxicity": {}.
    Currently, Toxicity prediction cannot be further configured with any parameters.
    If missing or null, no toxicity predictions will be generated.

get_model_type

Get the configuration model type.

@classmethod
def get_model_type(cls) -> ModelType

Returns:

  • ModelType - Model type.

FacemeshConfig

Configuration for the facemesh model.

@dataclass
class FacemeshConfig(ModelConfigBase["FacemeshConfig"])

get_model_type

Get the configuration model type.

@classmethod
def get_model_type(cls) -> ModelType

Returns:

  • ModelType - Model type.

ProsodyConfig

Configuration for the speech prosody model.

@dataclass
class ProsodyConfig(ModelConfigBase["ProsodyConfig"])

Arguments:

  • granularity Optional[str] - The granularity at which to generate predictions.
    Accepted values are word, sentence, utterance, or conversational_turn.
    The default is utterance.
    utterance corresponds to a natural pause or break in conversation
    conversational_turn corresponds to a change in speaker.
    This configuration is only available for the batch API.
  • identify_speakers Optional[bool] - Whether to return identifiers for speakers over time. If true,
    unique identifiers will be assigned to spoken words to differentiate different speakers. If false,
    all speakers will be tagged with an "unknown" ID.
    This configuration is only available for the batch API.
  • window Optional[Dict[str, float]] - Sliding window used to chunk audio.
    This dictionary input takes two entries: length and step representing
    the width of the window in seconds and the step size in seconds.
    This configuration is only available for the batch API.

get_model_type

Get the configuration model type.

@classmethod
def get_model_type(cls) -> ModelType

Returns:

  • ModelType - Model type.

BurstConfig

Configuration for the vocal burst model.

@dataclass
class BurstConfig(ModelConfigBase["BurstConfig"])

get_model_type

Get the configuration model type.

@classmethod
def get_model_type(cls) -> ModelType

Returns:

  • ModelType - Model type.

NerConfig (Batch API Only)

Configuration for the named-entity emotion model.

@dataclass
class NerConfig(ModelConfigBase["NerConfig"])

Arguments:

  • identify_speakers Optional[bool] - Whether to return identifiers for speakers over time. If true,
    unique identifiers will be assigned to spoken words to differentiate different speakers. If false,
    all speakers will be tagged with an "unknown" ID.
    This configuration is only available for the batch API.

get_model_type

Get the configuration model type.

@classmethod
def get_model_type(cls) -> ModelType

Returns:

  • ModelType - Model type.

Updated 3 months ago