ViT_B_16

Transformer based image classification model.

Image ClassificationCNNViT

Technical details

Model type

Image Classification

Train dataset

Imagenet

Test dataset

Imagenet

Hardware

GPU

CPU

Mobile

NVIDIA

L4

V100-SXM2-16GB

T4

Size (MB)

Original

330.285

CLIKA

85.456

Accuracy (top1)

Original

81.070

CLIKA

80.864

Speed (FPS)

Original

135.471

CLIKA

925.862