ModelVision Transformer Base

ViT-B

The base-scale Vision Transformer model, which splits images into patches and processes them with self-attention mechanisms; it is widely used in computer vision tasks.

Timeline (last 90 days)

Oct 27

On vision tasks, UTA also improves the performance of ResNet50 and ViT-B on ImageNet.

Unverified50%

Oct 20

At the full VTF900 gallery scale (N=900) under frontal conditions, the ViT-B model achieves 97.68% AUC and 51.10% TAR@1%FAR.

Unverified50%

All Facts (2)

Unverified

On vision tasks, UTA also improves the performance of ResNet50 and ViT-B on ImageNet.

50%Unverified

At the full VTF900 gallery scale (N=900) under frontal conditions, the ViT-B model achieves 97.68% AUC and 51.10% TAR@1%FAR.

50%

Source Articles

UTA Learning Rate Decay: Closed-Loop Adaptation Breaks Through Training BottlenecksOct 27 VTF900: A 900-Subject Visible-Thermal Face Recognition Benchmark ReleasedOct 20