Custom written python NNet to recognize dgits from images. NNet is trained on MNIST dataset and then tested on separate data. Accuracy ~85%, could be improved with tweaking and training on bigger data set.
- using MNIST dataset
- Neurons and layers
- L0 - 784 - X - INPUT layer
- L1 - 64
- L2 - 16
- L3 - 10 - Y - OUTPUT layer
Interface:
