[WIP] Implemention of Recurrent Models of Visual Attention

• Requirments

---

1. pytorch1.0+
2. python 3.6
3. opencv

• Structure

---

src

• model.py: Contains the main model

• dataloader.py: Dataloader for pytorch

• glimpse.py: Glimpse Sensor Network used in model

• utils.py: Utilities functions

• solver.py: Main training code

• Results

---

Results for small dataset of 100 images. Training Accuracy reached 100% indicating that the model is able to overfit a small dataset Training Accuracy reached 100% on validation set