Local Tandem Learning (LTL)

The official code release for the Local Tandem Learning (LTL) framework.

Usage

This repo contains the code for image classification on the Cifar10 dataset with the VGG11 and VGG16 architectures.

Step 1. Pre-train baseline ANN models

run ./examples(ann_to_snn)/ANN_baseline/cifar10_vgg11_base_model.py or cifar10_vgg16_base_model.py

Step 2. Transfer to SNN model

a. Using offline version:

run ./examples(ann_to_snn)/offline_LTL/cifar10_main_svgg11_offline.py or cifar10_main_svgg16_offline.py

b. Using online version:

run ./examples(ann_to_snn)/online_LTL/cifar10_main_svgg11_online.py or cifar10_main_svgg16_online.py

Note: Remember to change the 'home_dir' and 'data_dir' to the file path of your local machine