Skip to content

Latest commit

 

History

History
22 lines (15 loc) · 425 Bytes

README.md

File metadata and controls

22 lines (15 loc) · 425 Bytes

one-bit-quantization

Code for APSIPA 2019 paper

Requirements

  • sacred
  • pytorch setup

Training

python train-exp.py will train all 15 models

Test

You need to specify which model to load. The test will do both normal and adversarial test.

For example:

To test 1 bit model

python eval-exp.py with bit=1 dithering=Fasle

To test 1 bit dithered model

python eval-exp.py with bit=1 dithering=True