Table 2 Hyper-parameters of the proposed method.

From: Integrating simplified Swin-T with modified EFS-Net for attention-guided underwater pipelines segmentation in complex underwater environments

Hyper parameter

Description

Epochs no.

100

Train dataset

60% of dataset

Test dataset

20% of dataset

Validation dataset

20% of dataset

Learning algorithm

Adam

Learning rate

0.0001

Activation function

Rectified Linear Unit (RLU)

Batch normalization

16

Validation

k-fold (k = 5)

Loss function

(Cross Entropy)+(Dice Loss)

Back to article page