VideoChat2Options

Options & Configuration · AiDotNet.VisionLanguage.VideoLanguage

Configuration options for VideoChat2: progressive video training with diverse data.

Inheritance: ModelOptions → GenerativeVLMOptions → VideoLanguageOptions → VideoChat2Options

For Beginners

These options configure the VideoChat2 model. Default values follow the original paper settings.

Constructors

Constructor Summary
VideoChat2Options() Initializes a new instance with default values.
VideoChat2Options(VideoChat2Options other) Initializes a new instance by copying from another instance.

Properties

Property Summary
int NumQFormerHeads { get; set; } Number of attention heads inside the Q-Former (BLIP-2 default 12).
int NumQFormerLayers { get; set; } Number of Q-Former transformer blocks (BLIP-2 default 12).
int NumQueryTokens { get; set; } Number of learnable Q-Former query tokens that cross-attend to the video features (BLIP-2 default 32).
int QFormerDim { get; set; } Q-Former hidden width (BLIP-2 default 768).