validation sequences vs testing sequences

[Randy Chen]

How are the validation and testing sequences chosen in params.py? They do not align with the original paper at all. Moreover, the original paper does not mention how the validation dataset is chosen, what is indeed mentioned is the following:
"In order to have a
separate dataset for testing, only the Sequence 00, 02, 08
and 09 which are relatively long are used for training. The
trajectories are segmented to different lengths to generate
much data for training, producing 7410 samples in total. The
trained models are tested on the Sequence 03, 04, 05, 06, 07
and 10 for evaluation."

Can someone clarify? In params.py the training set was chosen as [00, 01 ,02 ,05 ,08, 09], and the testing data set was ['04', '05', '07', '10', '09'], why??

[Marco Chan]

It also confuses me as testing dataset should not be used in training and validation dataset.