Multi-Layer Capsule Network and its evaluation implemented by Vikram Shenoy, Alexander Chowdhury, and Harry Hartenstine.
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.
This model was trained on Google Colab. To reproduce the results, follow the following instructions
1) Go to CapsNet_AT&T/MLCN
2) Copy the main.py file onto a Google Colab Cell
3) Upload the make_graph.py, att_MLCN.py, and utils.py file onto Google Colab
4) Change the mode from "Train" to "Test", "VPA" (Capsule Vector Perturbation Analysis), and "LVG" (Latent Vector Generation)
The above instructions apply to the other codes as well with some minor changes.
Results can be found in the Network Comparisons folder. One such result is displayed below.
More details can be found in the Results section of our paper.
- PyTorch- Deep Learning Framework
- Google Colab - Cloud Service
- Vikram Shenoy - Vikram Shenoy
- Alexander Chowdhury - Alexander Chowdhury
- Harry Hartenstine - Harry Hartenstine