search

suriyadeepan / practical_seq2seq

A simple, minimal wrapper for tensorflow's seq2seq module, for experimenting with datasets rapidly
http://suriyadeepan.github.io/2016-12-31-practical-seq2seq/
GNU General Public License v3.0
569 stars 270 forks source link

ImportError: No module named datasets.cornell_corpus #10

Open EN10 opened 7 years ago

EN10 commented 7 years ago

This seems like it's a noob error, but after a bit of googling still couldn't get it to work. So I cloned the repo and tried running 04-Cornell-Movie-Dialog-Bot.py and I get the error: "ImportError: No module named datasets.cornell_corpus" I guess it's an issue with my path, I tried to add: export PYTHONPATH="${PYTHONPATH}:~/seq2seq/datasets/" to .bashrc but no luck This seemed to be the recommended action after reading many stackoverflow threads. Any advice?

suriyadeepan commented 7 years ago

Make sure you have datasets/cornell_corpus/data.py in your local folder. Get the raw data "movie_lines.txt" and "movie_conversations.txt" from here and put them under datasets/cornell_corpus/raw_data/. I should have added these instructions to README. Will do it now.

EN10 commented 7 years ago

Still have the same issue: I have downloaded the Cornell Movie--Dialogs Corpus zip and unzipped it

"ls datasets/cornell_corpus/" shows: README.txt data.py movie_lines.txt chameleons.pdf movie_characters_metadata.txt movie_titles_metadata.txt cornell_movie_dialogs_corpus.zip movie_conversations.txt raw_script_urls.txt

but I still get: "ImportError: No module named datasets.cornell_corpus"

jihopark commented 7 years ago

I think this is because u are using python 2 which needs init.py to make packages