Update README.md

README.md

@@ -3,6 +3,12 @@ tags:
 - model_hub_mixin
 - pytorch_model_hub_mixin
 ---
+# AG-Net
+This is an implementation (the first public implementation as far as I know) of AG-Net as described in the paper "Attend and Guide (AG-Net): A Keypoints-driven Attention-based Deep Network for Image Recognition" by  Asish Bera, Zachary Wharton, Yonghuai Liu, Nik Bessis, and Ardhendu Behera. I include the weights for the model which achieves 98.3% accuracy on the test data. 
+
+I include the iPython Notebook in my Github where I derived the model to "show my work", but I also include a few python files which contain the combined code from the notebook. The files and functions are the following: utils.py, which contains the data fetching and augmentations, models.py, which contains the torch modules, train.py, which trains the model and optionally saves it, and test.py, which tests the model.
+
+Note that I coded all of this on my own with the exception of the "Intra Self-Attention" module which I obtained from https://github.com/heykeetae/Self-Attention-GAN/blob/master/sagan_models.py (also note that there it's called Self-Attention (Self_Attn module)). 
 
 This model has been pushed to the Hub using the [PytorchModelHubMixin](https://huggingface.co/docs/huggingface_hub/package_reference/mixins#huggingface_hub.PyTorchModelHubMixin) integration:
 - Library: [More Information Needed]