Exploring Features of TF2.0 By Developing a Cat/Dog Classifier
This project in Deep Learning aims to explore features in Tensorflow 2.0 by developing Cat/Dog classifier and evaluating their performance. Kaggle's Dogs vs. Cats dataset is used for this project. Three different classifiers are built in this project and the model weights are available for download via the following link. The BuildingCatDogClassifierInTF2.ipynb Jupyter Notebook describes the entire project and the results/observations. Further, the efficiency of the developed models is evaluated by inferencing on random images of cats & dogs downloaded from the internet. The inference dataset has been uploaded to this repository and is avilable in the following folders:
- CatDogClassifierInferenceImages.zip : contains inference images of cats & dogs downloaded from the internet.
- CatDogTogetherInferenceImages.zip : this folder contains images where both cat & dog are present in the same image; these images are used to gain an intuition of the features which impact the developed models' predictions.