DNAS-Compression

Model compression techniques with differentiable neural architecture search.

Currently, pruning and quantization are supported.

• Pruning : Channel-level / Group-level
  • Sparsity and group size can be set
• Quantization : Uniform quantization
  • Bitwidth can be set

References

This project is implemented based on FBNet reproduced version.

• FBNet [https://github.com/AnnaAraslanova/FBNet.git][FBNet]

Usage

• Requirement:
  • pytorch 1.6.0
  • tensorboardX 1.8
  • scipy 1.4.0
  • numpy 1.19.4

1. Choose what type of compression would you run

   • Pruning ( channel / group )
   • Quantization
   • Both pruning and quantization

2. Edit hyperparmeters in supernet_functions/config_for_supernet.py

   • Usual hyperparameters
     • batch size
     • learning rate
     • epochs
   • Special hyperparameters (pay attention to it!)
     • alpha, beta for adjustment of flops loss
     • w_share_in_train
     • thetas_lr
     • train_thetas_from_the_epoch

3. Run supernet_main_file.py

   Quick start command:

   python3 supernet_main_file.py --train_or_sample train