Table 2 Training hyperparameters for experiments.

From: Deep multi-task learning framework for gastrointestinal lesion-aided diagnosis and severity estimation

Config

Value/parameter

Optimizer

AdamW

Activation

ReLU

Batch size

16

Learning rate

1.00e − 4

Learning rate schedule

StepLR

Dropout

0.1

Epochs

100

Weight decay

1e-6

Back to article page