FuseFormerOptions

Options & Configuration · AiDotNet.Video.Options

Configuration options for the FuseFormer video inpainting model.

Inheritance: ModelOptions → FuseFormerOptions

For Beginners

FuseFormer options configure the transformer video inpainting model.

How It Works

References:

  • Paper: “FuseFormer: Fusing Fine-Grained Information in Transformers for Video Inpainting” (Liu et al., ICCV 2021)

FuseFormer uses a transformer architecture with soft split and composition operations to fuse fine-grained spatial-temporal information at multiple scales for high-quality video inpainting with better detail preservation.

Constructors

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

Properties

Property Summary
OnnxModelOptions OnnxOptions { get; set; } ONNX runtime options for inference mode.
VideoModelVariant Variant { get; set; } Model variant controlling capacity and speed trade-off.
double DropoutRate { get; set; } Dropout rate for regularization.
double LearningRate { get; set; } Learning rate for training.
int NumFeatures { get; set; } Number of base feature channels.
int NumHeads { get; set; } Number of attention heads in the multi-head attention modules.
int NumTransformerLayers { get; set; } Number of transformer layers for temporal fusion.
int PatchSize { get; set; } Patch size for the soft split operation that divides features into overlapping patches.
string ModelPath { get; set; } Path to the ONNX model file for inference mode.