sharad461 / nepali-translator Goto Github PK

hello,

I was attempting to install the requirements described in the readme with python 3.11 but it throws a bunch of compatibility errors. What version of python is expected for this project? 3.7? 3.8? 3.9? which one?

also what version of torch is required?

What are the python, pytorch, fairseq versions that you used to run ?

Hello I couldn't find this in the readme, and I have a few questions.

1. are the train.en and train.ne files in the collected folder of the corpus after or before data cleaning?
2. which of the corpus was used in the paper? corpus.zip or corpus_NOV2019.tar.gz? both seem to have fewer sentences than reported in the paper. For the second one, the files inside collected have 1780248 sentences whereas the paper mentions about 1810000 sentences.

Thanks.

I have been trying to convert your source data from .en or .np file into .csv file without losing the format. I tried to do it by converting into excel and eventually into .csv file but the format gets all over the place. Any tips? Please let me know. Thank you

AttributeError arising from this line specifically:

gen = interactive.Generator("app/ne_en_bpe2000", "app/models/ne_en.pt")

The error looks like this on flask server.