Skip to content

Latest commit

 

History

History
20 lines (12 loc) · 875 Bytes

README.md

File metadata and controls

20 lines (12 loc) · 875 Bytes

pruning_lenet300-100

Pruned LeNet300-100 using Iterative pruning based on https://arxiv.org/abs/1506.02626

Steps to follow:

  1. create conda enviornment with tensorflow>2.0 and matplotlib.

  2. Run main.py and update quality_parameter value to change pruning theshold. (Iterative training (prune+retrain) in single iteration and run until we get pruning accuracy >= original model accuracy. Threshold is calculated based on standard deviation of weights and sensitivity/quality factor).

  3. Results are uploaded in results folder. original model, pruned model, and retarin model accuracy & compression rate during iteraive pruning.

![Pruning Results] Pruning Results Pruning Results