Table 1 Hyperparameter tuning.

From: Artificial intelligence-powered prediction of AIM-2 inflammasome sequences using transformers and graph attention networks in periodontal inflammation

Cuda

TRUE

TRUE2

TRUE3

Seed

43

43

43

num_workers

4

4

4

num_class

2

2

2

Kmer

3

3

3

heatmap_seq

save_figure_type

png

png

png

Mode

train-test

train-test

train-test

Type

prot

prot

prot

Model

TRANSFORMERS

GAT

RNN-CNN

datatype

userprovide

userprovide

userprovide

interval_log

10

10

10

interval_valid

1

1

1

interval_test

1

1

1

Epoch

50

50

50

Optimizer

Adam

Adam

Adam

loss_func

C.E.

C.E.

C.E.

batch_size

4

8

32

L.R.

1.00E-05

0.0001

0.0001

Reg

0.0025

0.0025

0.0025

Gamma

2

2

2

Alpha

0.25

0.25

0.25

max_len

35

207

52

dim_embedding

32

32

32

minimode

modelCompare

modelCompare

modelCompare

if_use_FL

0

0

0

if_data_aug

0

0

0

if_data_enh

0

0

0

CDHit

[‘1’]

[‘1’]

[‘1’]