outstandingcandy/dssm

Deep structured semantic model

dssm's Issues

Cannot compile success, need help

Hi, I try to do DSSM experiment recently, however, I stucked in your code. Would you plz help me give example data to run success? Thanks a lot!

Maybe there is a bug in generate_sparse_training_data

You calculate the doc_size with the fomula as:
doc_size = len(positive_doc_feature_list)
and select the doc_id with the fomula as:
negative_sample_id = random.randint(0, doc_size-1)
but the negative_sample_id is beyond the range of doc_id number, when you do the judging:
if negative_sample_id not in query_doc_dict[query]:
since the number of doc_id is equal with the number of queries