Week 1 Homework Question

Course Q&A Natural Language Processing NLP with Sequence Models
1

Hello,

I have a question about a portion of the codes in the HW C3_W1_Assignment. In Exercise 02 Implement Data Generator, there are these lines:

    if stop:
        break;

    # Update the start index for positive data 
    # so that it's n_to_take positions after the current pos_index
    pos_index += n_to_take
    
    # Update the start index for negative data 
    # so that it's n_to_take positions after the current neg_index
    neg_index += n_to_take
    
    # Get the max tweet length (the length of the longest tweet) 
    # (you will pad all shorter tweets to have this length)
    max_len = max([len(t) for t in batch])

My question is, why update pos_index and neg_index as such above? The reason I thought might be that this prevents not having enough batch samples towards the end of data_pos and data_neg cycle. But if this is the case, then it means the algorithm is skipping data samples in the middle too.

Thanks,
John

2

This issue is addressed in HW2 (week 2) i.e., the codes above are not used.