Data2Vec2Options
Options & Configuration · AiDotNet.Audio.Foundations
Configuration options for the data2vec 2.0 self-supervised audio foundation model.
Inheritance: ModelOptions → Data2Vec2Options
For Beginners
data2vec 2.0 is a foundation model that learns by predicting its own hidden representations - like studying by explaining things to yourself. Unlike HuBERT which predicts discrete labels, data2vec predicts rich continuous features. Version 2.0 is much faster to train while maintaining quality. It works for audio, images, and text.
How It Works
data2vec 2.0 (Baevski et al., 2023, Meta) is a self-supervised learning framework that predicts contextualized latent representations rather than modality-specific targets. Version 2.0 is 16x faster than v1 through efficient data encoding and a novel self-distillation objective. It achieves strong results on speech, vision, and language tasks with the same method.
Constructors
| Constructor | Summary |
|---|---|
Data2Vec2Options() |
Properties
| Property | Summary |
|---|---|
OnnxModelOptions OnnxOptions { get; set; } |
Gets or sets the ONNX runtime options. |
double DropoutRate { get; set; } |
Gets or sets the dropout rate. |
double EMADecay { get; set; } |
Gets or sets the teacher EMA decay rate. |
double LearningRate { get; set; } |
Gets or sets the learning rate. |
double MaskProbability { get; set; } |
Gets or sets the masking probability for training. |
int FeedForwardDim { get; set; } |
Gets or sets the feed-forward inner dimension. |
int HiddenDim { get; set; } |
Gets or sets the model hidden dimension. |
int MaskSpanLength { get; set; } |
Gets or sets the mask span length. |
int NumHeads { get; set; } |
Gets or sets the number of attention heads. |
int NumLayers { get; set; } |
Gets or sets the number of transformer layers. |
int SampleRate { get; set; } |
Gets or sets the audio sample rate in Hz. |
int TopKLayers { get; set; } |
Gets or sets the number of top-k layers averaged for teacher targets. |
int[] ConvChannels { get; set; } |
Gets or sets the convolutional feature extractor channels. |
int[] ConvKernels { get; set; } |
Gets or sets the convolutional feature extractor kernel sizes. |
string ModelPath { get; set; } |
Gets or sets the path to the ONNX model file. |
string Variant { get; set; } |
Gets or sets the model variant (“base” or “large”). |