Algorithm 1. Channel L1 norm-based pruning algorithm

Input: Training data: X, pre-trained weights: W, hyperparameter group,

pruning strategy:

neural network training parameters (such as learning rate, batch size, etc.),

Output: Compressed weights

1: for pruning strategy do

2:   Obtain θn based on W and hyperparameter group

3:   if Wn < θn do

4:     Channel L1 norm-based pruning on Wn

5:   else do

6:     Retrain the network

7: end for