This an image classifier that tries to predict whether the selected image is a Dog or a Cat.
Select a image:
This model was trained with Keras library in python with Tensorflow as backend.
I used the "Dogs vs Cats Redux: Kernels Edition" dataset un Kaggle.
I trained it using convolutional neural network with a data augmentation technique based on my readings in François Chollet's book Deep Learning with Python.
You can see all the data preparation, training, evaluation, test, accuracy in the Kaggle notebook: https://www.kaggle.com/keogh24/dogs-vs-cats-keras-data-augmentation/
I saved the model in HDF5 format then I exported to tensorflow.js, loaded it in a gatsby.js static website.
This will only try to predict whether the image is a dog or cat no matter what image you upload.
This is all done in the web browser using tensorflow.js
This was a site I created for a tech talk I gave in Feb 27 2020. Here are the slides.