Readme.md and add-ons

Trained layers (theta1 and theta2) using keras_dataset

Minnor changes

Change the input to be a Grayscale and match to the dataset

Change Input Size 20x20 -> 28x28 (px)

Output Layer function change sigmoid -> softmax

First Layer and Hidden Layer activation function change sigmoid -> ReLU

Use saved data (from nn.json), and bug (dir) fixed)

Merge branch 'main' of https://github.com/TiagoSBittencourt/Basic-OCR-AI-

Initial project upload, (AI not trained, yet)

Update README.md

Initial commit