MNIST is a basic classification dataset that most of the people use to kick-start their image classification journey. It consists of 10 classes(10 digits) in the handwritten format and the main task is to classify them correctly. It consists of 70000 images each of size 28x28.