search

KhalilMrini / LAL-Parser

Neural Adobe-UCSD Parser, the current State of the Art in Constituency and Dependency Parsing.
138 stars 24 forks source link

Fix some minor bugs and add instructions #2

Closed yangky11 closed 4 years ago

yangky11 commented 4 years ago

Hi,

Thanks for releasing the code. I tried it and was able to reproduce the exact numbers. I fixed some minor issues along the way, which could be useful.

  1. The pre-compiled EVALB/evalb binary does not work on my machine and I had to re-compile it. So I removed the file from the repo and added instructions in the README.
  2. The uint8 in src_joint/KM_parser.py does not work for newer versions of PyTorch. This is a known issue originated from self-attentive-parser.
  3. The default eval-batch-size does not fit into GPUs with 11~12GB memory. So I modified test.sh
KhalilMrini commented 4 years ago

Hi Kaiyu,

Thank you so much for trying our code!

Your feedback is very useful, I am merging your suggestions. :-)

Please let me know if you come across something else.