ML Image Recognition Demo
The application is an end-to-end web-based solution for image classification using the CIFAR-10 dataset.
It consists of a frontend built with HTML, CSS, and JavaScript, as well as a backend developed using Python, Flask, and PyTorch.
Classify an image
Please upload an image from the CIFAR-10 dataset,
from the test folder in order to test the output of the model.
The Custom CNN is a rudimentary model, just to prove the concept, so please expect modest results. 😊
Re-train the model
A model already exists and is ready to be used, but you can also re-train it. Using a higher number of epochs will take longer to compute, but will not necessarily increase prediction accuracy. For resource-consciousness reasons, you can only train the Custom module.