Table 4 Pre-trained and TL-based feed-forward NN architectures for the GDL geometric tortuosity prediction.

From: Transfer learning assessment of small datasets relating manufacturing parameters with electrochemical energy cell component properties

Neural Network

Number of hidden layers

Number of nodes per layer

Total number of trainable parameters

Pre-trained

3

3

46

TL-based

1 (plus the frozen previous layers)

3

10

Back to article page