Wav2Vec2LidOptions
Options & Configuration · AiDotNet.Audio.LanguageIdentification
Configuration options specific to Wav2Vec2 language identification.
Inheritance: ModelOptions → LanguageIdentifierOptions → Wav2Vec2LidOptions
Constructors
| Constructor | Summary |
|---|---|
Wav2Vec2LidOptions() |
Properties
| Property | Summary |
|---|---|
double FeatureProjectionDropout { get; set; } |
Gets or sets the feature projection dropout rate. |
double HiddenDropout { get; set; } |
Gets or sets the hidden dropout rate. |
int ClassifierProjectionSize { get; set; } |
Gets or sets the width of the projector before mean pooling (transformers’ classifier_proj_size, 256). |
int FeatureEncoderDim { get; set; } |
Gets or sets the width of each feature-encoder stage. |
int HiddenSize { get; set; } |
Gets or sets the hidden size of the transformer. |
int IntermediateSize { get; set; } |
Gets or sets the intermediate size for feed-forward layers. |
int NumAttentionHeads { get; set; } |
Gets or sets the number of attention heads. |
int NumLayers { get; set; } |
Gets or sets the number of transformer layers. |
int PositionalConvGroups { get; set; } |
Gets or sets the group count of the convolutional relative positional embedding. |
int PositionalConvKernel { get; set; } |
Gets or sets the kernel width of the convolutional relative positional embedding. |
int[] FeatureEncoderKernels { get; set; } |
Gets or sets the kernel width of each feature-encoder convolution over the raw waveform. |
int[] FeatureEncoderStrides { get; set; } |
Gets or sets the stride of each feature-encoder convolution, aligned with FeatureEncoderKernels. |