Skip to main content
. 2022 Jan 27;22(3):991. doi: 10.3390/s22030991
Algorithm 1 Training strategy
  • 1:

    lr=2e−4; optimizer=Adam(lr,beta_1=0.5);

  • 2:

    α=1; β=1;

  • 3:

    for epoch in epochs do

  • 4:

        if epoch>(epochs × 65%) then

  • 5:

            α=1, β=0.5;

  • 6:

            lr+=−5e−8;

  • 7:

        end if

  • 8:

    end for

  • 9:

    α=1, β=1;