tuple_extract.ipynb contains code that extract image tuple from video.
tupledata.py and clipdata.py are implement of pytorch dataset for image tuple classification task and clip ordering task respectively.
Net.py and ViT_var.py contains definition of our model.
show_model.py are used to check the architecture of the model.
train.py and train_vit.py are training process for the two tasks.
pair_wise_inference.py are used to evaluate our models.