This is a from the scratch python implementation of single layered neural network. The project works for python versions 2.* or 3.*
MNIST dataset is considered for evaluation of the model.
directory:\Codes> pip install requirements.txt
directory:\Codes> python main.py --help
usage: main.py [-h] [-d DATA] [-l LABEL] [-p PLOT] [-s SAVE] [-ts TEST_SIZE] [-e N_EPOCH] [-lr LR] [-hn HIDDEN_NEURONES]
Implementing neural network without any ML libraries
optional arguments:
-h, --help show this help message and exit
-d DATA, --data DATA Dataset path
-l LABEL, --label LABEL
Dataset path
-p PLOT, --plot PLOT Plot path
-s SAVE, --save SAVE Save path
-ts TEST_SIZE, --test_size TEST_SIZE
Test size
-e N_EPOCH, --n_epoch N_EPOCH
Number of epochs
-lr LR, --lr LR learning rate
-hn HIDDEN_NEURONES, --hidden_neurones HIDDEN_NEURONES
Neurones in the hidden layer
directory:\Codes> python main.py