search

nitin-rathi / hybrid-snn-conversion

Training spiking networks with hybrid ann-snn conversion and spike-based backpropagation
https://openreview.net/forum?id=B1xSperKvH
97 stars 24 forks source link

Retraining the ANN generated model #1

Open ksouvik52 opened 4 years ago

ksouvik52 commented 4 years ago

Hi I have been trying to reproduce the results by training an ANN from scratch with the same network model of VGG16 as described in your code. Then I tried to use the main.py to convert my saved model .pt to SNN. I am facing following issues:

  1. My ANN training with the model given in self_models/ gives test accuracy around 91% , which is around 1.81% less that what is reported in the paper.
  2. When I use this saved ANN model to convert to SNN domain, the accuracy I get after 20 epochs is around 7% less that the ANN baseline.

Please help me with the issue to reproduce the results.

nitin-rathi commented 4 years ago

I have updated the repository with trained models. I have also listed some reasons why you may be noticing the accuracy drop. I have also included trained SNN models which should give you the reported accuracy. Hope this helps!

ksouvik52 commented 4 years ago

Hi Nitin, RESNET_SNN_STDB class is missing in the spiking models file. Can you include that in the file?

kugatsu-sudo commented 3 years ago

I am very curious, why the accuracy of running snn.py alone is even better than running ann.py first and then running snn.py

kugatsu-sudo commented 3 years ago

Have you tried running snn.py alone