https://www.mdpi.com/2075-4418/14/13/1439
- download and rename void-movienet_50f directory as movienet_50f
- download and rename void-npef_movienet_50f directory as npef_movienet_50f
- delete "delete_me_after loading _videos.txt" files in last-level subdirectories
- download and use "create_dataset" notebooks
- a csv file per notebook will be created: take care to give it a suitable experiment name
- download and open movinet_loss_curves.xlsx
- duplicate one of the folder, give it the experiment name
- copy/paste the printed training output in the first cell, and that's all!
- download echocg_video_metrics_and_graphs notebook
- modifying the adhoc cells, downlad as pandas dataframes the csv files that were created by traning notebooks
- adapt the model dictionnaries to take into account your choice of experiments
- run remaining cells