The question for training process

[hit-lu]

Dear Authors,
We note that there seems to be a problem with the processing of the training data in file "LSTM.py". The data "X_train" includes the information of heuristic choice (Is this an intentional design?). We have also noticed inconsistencies in the sequence data. May I ask if it is a problem. Thanks for reply.

A example of the traing data:
your design （2 x 107 elements）
‘’‘
feature code (7 elements) + problem code (100 elements)
heuristic code (7 elements, which have been included in the training data by mistake) + feature code (7 elements) + problem code (only the first 93 elements)
’‘’

the right design we think
‘’‘
feature code (7 elements) + problem code (100 elements)
feature code (7 elements) + problem code (100 elements)
’‘’