salt-nlp / formalitystyletransfer Goto Github PK

I want to know how to solve this probelm.

Hi, thanks for sharing your code with us.

I wonder what the SPLIT "plain" means in preprocess.sh:

https://github.com/GT-SALT/FormalityStyleTransfer/blob/a86d287d0c48238f7cd39f6f34b465b0b7ccb2f4/preprocess.sh#L6

https://github.com/GT-SALT/FormalityStyleTransfer/blob/a86d287d0c48238f7cd39f6f34b465b0b7ccb2f4/preprocess.sh#L26

Also, the preprocess.sh script requires files named like "train.source" and "val.target". The GYAFC dataset does not follow this format, however. Should I reformat the original files into the required format? In that case, which files should be renamed "plain.source" and "plain.target"?

Hi,

Could you please provide the file language_pair_dataset.py for data preprocessing? Thank you.

in this module "fairseq/data/iterators.py " in line 49 there is an error named "StopIteration" which is not compatible with updated Faiseq. it means there is a conflict with fairseq folder on your repository with pytorch/fairseq. how would you run this code? please write a clear description about the process of runnig the code