This repo regroups the results of this research paper: https://arxiv.org/abs/1912.03467
Multiple training configurations were tested to understand the impact of each parameter on the Attention model.
To visualize the results of each training using Tensorboard :
- Open the Anaconda Prompt
- Change dir to the root of the project
- Tensorboard --logdir logs/
- Open this link http://localhost:6006/
Details in ./environment.yml
The original source code was provided by https://github.com/tianyu-tristan/Visual-Attention-Model/tree/master/MNIST/ram