Test

Testing script is ./pipeline.py (the same as ./run.py)
--img_dir argument need to be specified (path to the directory, containing images)

The test script can work both on gpu and cpu:

Weights

Download 'model_places2.pth' weights from here: https://drive.google.com/open?id=1SGJ_Z9kpchdnZ3Qwwf4HnN-Cq-AeK7vH and place it inside ./DFNet/model/ directory

Training script is ./DFNet/train_vlad.py, is based on this: https://github.com/Yukariin/DFNet/blob/master/train.py
The model was trained on Places365-Standard high-resolution images: http://places2.csail.mit.edu/download.html
To train the model you need to create training and validaton .flist files, containing the paths to images
Example of how to generate your own .flist files is in ./DFNet/flist.py script

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
DFNet		DFNet
result		result
.gitignore		.gitignore
.gitlab-ci.yml		.gitlab-ci.yml
README.md		README.md
__init__.py		__init__.py
classical.py		classical.py
classical_biharmonic.py		classical_biharmonic.py
compare.py		compare.py
compare_test.py		compare_test.py
generate_masks.py		generate_masks.py
ohe_images_pix2pix.py		ohe_images_pix2pix.py
parse_txt.py		parse_txt.py
perimeter_draft.py		perimeter_draft.py
pipeline (copy).py		pipeline (copy).py
pipeline.py		pipeline.py
pipeline_3315.py		pipeline_3315.py
pipeline_dfnet.py		pipeline_dfnet.py
pipeline_dfnet_postprocess.py		pipeline_dfnet_postprocess.py
pipeline_dfnet_postprocess_old.py		pipeline_dfnet_postprocess_old.py
pipeline_dfnet_postprocess_rafa.py		pipeline_dfnet_postprocess_rafa.py
pipeline_dfnet_postprocess_vlad.py		pipeline_dfnet_postprocess_vlad.py
pipeline_generative_inpainting.py		pipeline_generative_inpainting.py
pipeline_mask.py		pipeline_mask.py
pipeline_train.py		pipeline_train.py
pipeline_train_black.py		pipeline_train_black.py
pipeline_train_pad.py		pipeline_train_pad.py
pipeline_train_pad_rafa.py		pipeline_train_pad_rafa.py
poissonblending.py		poissonblending.py
postprocessing.py		postprocessing.py
predict256.py		predict256.py
run.py		run.py
test_final.txt		test_final.txt
test_wavelets.ipynb		test_wavelets.ipynb
train_dict.pkl		train_dict.pkl
trypoissonblending.py		trypoissonblending.py