EmuEditOptions

Options & Configuration · AiDotNet.VisionLanguage.Editing

Configuration options for Emu Edit: precise image editing via recognition and generation tasks.

Inheritance: ModelOptions → GenerativeVLMOptions → EditingVLMOptions → EmuEditOptions

For Beginners

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

Constructors

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

Properties

Property Summary
bool EnablePreciseEditing { get; set; } Gets or sets whether to use recognition-guided precise editing.
int CrossAttentionDim { get; set; } Gets or sets the cross-attention width shared by the U-Net’s conditioning projections and the edit head.
int UNetBaseChannels { get; set; } Gets or sets the base channel width of the denoising U-Net.
int UNetNumResBlocks { get; set; } Gets or sets the number of residual blocks per U-Net level.
int VaeBaseChannels { get; set; } Gets or sets the base channel width of the VAE.
int VaeNumResBlocksPerLevel { get; set; } Gets or sets the number of residual blocks per VAE level.
int[] UNetAttentionResolutions { get; set; } Gets or sets the spatial resolutions at which the U-Net applies self-attention.
int[] UNetChannelMultipliers { get; set; } Gets or sets the per-level channel multipliers of the denoising U-Net.
int[] VaeChannelMultipliers { get; set; } Gets or sets the per-level channel multipliers of the VAE.