nn-runtime-earlyex

• A python based runtime tool for neural networks to achieve fast results using early exits

1. Distance based Early Exit Framework(DEX)

DEX is a framework that tries to insert distance based early exit classifiers.

Train backbone model

1. Run the following code to train backbone model:

python train_backbone.py

Train Cross Entropy

1. Run the following code:

python train_ce_branch.py

Train Metric based Classifier

python train_me_branch.py

2. LazyNet: Lazy entry into Neural Networks(ICTC2022)

Usage:

python 1_train.py

base.yml is based on yaml, and contains config files

Acknowledgement

This work was supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government (MSIT) (No. 2021-0-00907, Development of Adaptive and Lightweight Edge-Collaborative Analysis Technology for Enabling Proactively Immediate Response and Rapid Learning).