IBM / Graph2Tree

Graph4Tree is a simple example code for our EMNLP'20 Findings paper idea.
Apache License 2.0
26 stars 9 forks source link

Reading the parsing file #4

Closed Juanjdurillo closed 3 years ago

Juanjdurillo commented 3 years ago

Hi there! Thanks for making the software available. After reading the phenomenal paper, I was interesting in the code and set to do some experiments, but I am having troubles in making it run (maybe simply my misunderstanding.)

Running the train.sh fails for me at https://github.com/IBM/Graph2Tree/blob/04e5b96134816037aa0dfb41e39fa4f9d60f099d/src/graph2tree.py#L337 as the file map.pkl is not found. Using the file file_for_parsing.pkl which is the only pkl file in the repo, I get the following error though:

 word_manager, form_manager = managers
ValueError: too many values to unpack (expected 2)

I am sure I am missing something here (should I generate the map.pkl in a different way?) It would really great if you could help me out :). Thanks!

schenglee commented 3 years ago

Hi juanjo,

maybe you should run the graph construction scripts first to generate the file map.pkl: constituency.py.

Hope this can help you.

Juanjdurillo commented 3 years ago

Hi! Thanks as lot! It does help a lot :). I am not sure why it did not run before (as is part of the train.sh script)...