google/vit-large-patch16-224

The Vision Transformer (ViT) is a transformer encoder model (BERT-like) pretrained on a large collection of images in a supervised fashion . The model was fine-tuned on ImageNet-21k, a dataset consisting of 14 million images and 21k classes, at a resolution of 224x224 pixels . Images are presented to the model as a sequence of fixed-size patches (resolution 16x16), which are linearly embedded. One typically places a linear layer on top of the [CLS] token, as the last hidden state of this token can be seen as a representation of an entire image. By pre-training the model, it learns an,