ColorRefineNet
==============================================
In Shallow-UWNet and the proposed method;
Firstly, create account in wandb.ai Weights and Biases (W&B) which is a Track machine learning work.
TRAIN_DATABASE = EUVP
TEST_DATABASE = EU_Dark, UFO, UIEB

(1) Model Training
==================
python OrgTrain.py

***
To produce the snapshot checkpoint data, employ the folder designated as snapshots_folder.
input_images_path and label_images_path refer to the underwater images and their respective ground truth images for training data.
For validation data, test_images_path and GTr_test_images_path represent the underwater images and their corresponding ground truth images,
while output_images_path denotes the resulting output for the validation images.
***

(2) Model Testing
=================
python OrgTest.py
***
The snapshot_path has been specifically designated for test images.
The test_images_path and label_images_path parameters are set to ./data/TEST_DATABASE/input and
./data/TEST_DATABASE/label, respectively.
Ensure to specify the output_images_path folder; for instance, ./results/OUPUT_IMAGES.
***

(3) Compute the number of parameters
====================================
python compute_params.py

** Modify the file names for DLPFTest, BLPFTest, GLPFTest, and SLPFTest to different requirements.


(4) Dataset information
========================
Training dataset = EUVP (Train and Validation dataset)
Testing datasets = EU_Dark, UFO, UIEB