Models / timm / ViT-Base (Patch16, 224px, AugReg, IN21k)

ViT-Base (Patch16, 224px, AugReg, IN21k)

JSON →
vision

A Vision Transformer base model with patch size 16, pretrained on ImageNet-21k with AugReg for general-purpose image representation.

imagevision
homepage ↗docs ↗
Specs
Context & limits
—
context window
—
max output
Lifecycle
Dates
released2021-10-01
Resources
homepagehttps://huggingface.co/timm/vit_base_patch16_224.augreg_in21k ↗
docshttps://huggingface.co/docs/timm/en/models/vit ↗