# A conceptual breakdown of the data we have

#shape: [num_batches x batch_size x seq_length]
batch_tokens = [
[101, 1100, 87345, 1332,  102, 4589, 988, 874, 13598, 1332, 0,    0,    0],
[101, 987,  1332,  87345, 873, 4589, 102, 874, 13598, 1332, 1399, 1324, 1246],
...
]

#shape: [num_batches x batch_size x seq_length]
batch_location = [
[0,0,0,0,1,1,1,1,1,1,1,1,1],
[0,0,0,0,0,0,1,1,1,1,1,1,1],
...
]

#shape: [num_batches x batch_size]
batch_is_postive_lables = [
1,
0
]