This package runs training process for CNN in PyTorch.
Inputs:
dataset/
...
train_set/
test_set/
Outputs:
saved_models/
cnn1/
cnn2/
...
In order to run:
python3 main.py dataset/training_set/ dataset/training_set/train.csv dataset/test_set/ dataset/test_set/test.csv config/cnn1.yaml --device cuda
usage: main.py [-h] [--device DEVICE] src_train_dir src_train_csv src_test_dir src_test_csv config
positional arguments:
src_train_dir [input] path to train source directory
src_train_csv [input] path to input file with train annotations
src_test_dir [input] path to test source directory
src_test_csv [input] path to input file with test annotations
config [input] path to config yaml file
optional arguments:
-h, --help show this help message and exit
--device DEVICE [input] cpu or cuda