Skip to main content
. 2021 Aug 4;7(8):131. doi: 10.3390/jimaging7080131

Table 5.

Comparison of computational complexity and performance of DL models.

Model Name Number of Parameters Size on Disk Inference Times/Dataset Training Times/Dataset
Trainable Non-Trainable CPU (sec) GPU (sec) GPU (days)
C-ENET 793,917 11426 11 MB 16.857 4.026 4.22
ENET 363,069 8354 5.8 MB 12.833 3.505 3.47
ERFNET 2,056,440 0 25.3 MB 10.630 2.614 2.87