Table 1 Parameter values

From: Solving two-stage stochastic route-planning problem in milliseconds via end-to-end deep learning

  Parameter Value
Hyper parameters Epoch 8
Batch size 256
Training steps (per epoch) 2500
Optimizer Adam
Learning rate 0.003
Learning rate decay 0.95
Parameters of the MSDN d1 128
  d2 128
  d3 128
  |H| 3
Parameters of the IG algorithm α 3
  T0 500
  gmax 200
  c 0.95
  Sampling number 10,000