src/code/training.py

import torch
import sys
from fastai2.vision.all import *
from torchvision.utils import save_image


def get_y_fn(x):
    y = str(x.absolute()).replace('.jpg', '_depth.png')
    y = Path(y)

    return y


def create_data(data_path):
    fnames = get_files(data_path/'train', extensions='.jpg')
    data = SegmentationDataLoaders.from_label_func(data_path/'train', fnames=fnames, label_func=get_y_fn)
    return data


def train(data):
    data.num_workers = 0
    data.bs = 1
    learner = unet_learner(data, resnet34, metrics=rmse, wd=1e-2, n_out=1, loss_func=torch.nn.SmoothL1Loss())
    learner.fit_one_cycle(1, 1e-3)


if __name__ == "__main__":
    if len(sys.argv) < 3:
        print("usage: %s <data_path> <out_folder>" % sys.argv[0], file=sys.stderr)
        sys.exit(0)

    data = create_data(Path(sys.argv[1]))
    data.batch_size = 1
    data.num_workers = 0
    learner = train(data)

    learner.save(sys.argv[2])
    learner.show_results()