CNN Guide with Keras — Acc 0.997 (with Code)

CNN (Convolutional Neural Network) for digits recognition trained on the MNIST dataset. This is a 5 layered sequential network build on Keras API (Tensor flow backend). Firstly, I will prepare the data (handwritten digits images) then I will focus on the CNN modeling and evaluation.

I achieved 99.671% accuracy with this CNN trained in 1h30 on a Dual CPU (i5 2500k). For those who have a >= 3.0 GPU capabilities (from GTX 1650 — to recent GPUs), you can use TensorFlow-GPU with Keras. Computation will be much faster !!!

This Notebook follows three main parts:

• The data preparation
• The CNN modeling and evaluation
• The results prediction and submission

Data Preparation

Load data