Table 5 The average number of images processed by our models in one second at training and inference time.

From: In-domain versus out-of-domain transfer learning in plankton image classification

Model

BEiT

ViT

SWIN

ConvNeXt

Ensemble

Training (imgs/s) \(\uparrow \)

20.32

21.72

32.57

13.16

4.95

Inference (imgs/s) \(\uparrow \)

65.68

70.26

102.70

52.88

17.21

  1. The values have been evaluated based on 1000 iterations. The higher the value, the faster the processing time.
Back to article page