Genetics Classification Using Neural Networks

Pytorch Implementation for Genetics Classification. For further explanation please read my blog: Bridging the Gap Between Genetics and Neural Networks

Usage

Step 1: Generate the Embedding Matrix (optional) python utils_helpers.py

Step 2: Train the Network python run_tests.py -file_name your_file_name.pkl -batch_size 64 -n_epochs 1000 -use_embed_layer 0 -fold 1 -patience 100 -hidden_sizes 50 -dropout_1 0.8 -dropout_2 0.5

Parameters

Name	Required	Type	Description
file_name	required	str	bases and labels
batch_size	required	int	batch sizes [32, 64, 128, 256]
n_epochs	required	int	number of epochs
use_embed_layer	required	int	using the auxiliary network
fold	optional	int	fold of the embedding dataset
patience	optional	int	patience of the early stopping
hidden_sizes	required	int	hidden layer unit sizes
dropout_1	required	float	dropout of the first hidden layer
dropout_2	required	float	dropout of the second hidden layer

Requirements

Must

• Torch
• Numpy

Optional

• Matplotlib
• Pickle