OscarOptions

Options & Configuration · AiDotNet.VisionLanguage.Foundational

Configuration options for Oscar (Object-Semantics Aligned pre-training).

Inheritance: ModelOptions → FoundationalVLMOptions → OscarOptions

How It Works

Oscar (Li et al., ECCV 2020) uses detected object tags as “anchor points” to align image regions with text tokens, forming triples of (word tokens, object tags, region features) that are fed into a single BERT transformer.

Constructors

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

Properties

Property Summary
double ContrastiveLossWeight { get; set; } Gets or sets the contrastive loss weight for tag-text alignment.
int MaxImageRegions { get; set; } Gets or sets the maximum number of image regions.
int MaxObjectTags { get; set; } Gets or sets the maximum number of object tags per image.