Skip to main content
. 2022 Sep;12(9):4512–4528. doi: 10.21037/qims-22-33

Table 10. Analysis of the computation complexity.

Method Trainable parameters (M) GFLOPs Inference time (s/image) Training time (h)
MT 5.97 7.77 0.44 11.5
DT 16.37 9.03 0.13 7.25

Training time is the average of the training time spent on the 2 datasets. GFLOPs, Giga Floating point Operations Per Second; MT, medical transformer; DT, dilated transformer.