Closed joseph12346 closed 5 years ago
Here is an example script to generate the training code. File captions_train2014.json
and captions_val2014.json
can be obtained from MSCOCO dataset (see http://cocodataset.org/ )
https://gist.github.com/iamaaditya/bcae0a54b250e62c3be7e78f61de10df
How to create (neural-paraphrase-generation-dev\data\mscoco) train_vocab.txt. THX !