VisionTransformer¶

模型构建器¶

可以使用以下模型构建器来实例化 VisionTransformer 模型，可选择是否包含预训练权重。所有模型构建器内部都依赖于 torchvision.models.vision_transformer.VisionTransformer 基类。有关此类的更多详细信息，请参阅源代码。

`vit_b_16`(*[, weights, progress])	构建来自 An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale 的 vit_b_16 架构。
`vit_b_32`(*[, weights, progress])	构建来自 An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale 的 vit_b_32 架构。
`vit_l_16`(*[, weights, progress])	构建来自 An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale 的 vit_l_16 架构。
`vit_l_32`(*[, weights, progress])	构建来自 An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale 的 vit_l_32 架构。
`vit_h_14`(*[, weights, progress])	构建来自 An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale 的 vit_h_14 架构。