baolef/liif-3d

LIIF-3d

The code is based on Yinbo Chen's liif.

This is 3D implementation of the LIIF model.

Training

To train the model, you can run train.sh in the scripts folder

sh scripts/train.sh

To visualize the training results (eg. loss, training output), you can run tensorboard.sh

sh tensorboard.sh

If tensorboard.sh is run on a server, you can run local.sh on your local machine to access the visualization

sh local.sh

Testing

To test the model, you need to run test-div2k.sh in the scripts folder

sh scripts/test-div2k.sh

To summarize the testing results, you need to run summary.py

python summary.py

To run the baseline (nearest neighbour), you need to run baseline.py

python baseline.py

Demo

To directly calculate the output without computing metrics, you need to run demo.sh in the scripts folder

sh scripts/demo.sh