search

erickrf / nlpnet

A neural network architecture for NLP tasks, using cython for fast performance. Currently, it can perform POS tagging, SRL and dependency parsing.
http://nilc.icmc.usp.br/nlpnet/
MIT License
407 stars 104 forks source link

How to train new SRL model ? #23

Closed GraphGrailAi closed 8 years ago

GraphGrailAi commented 8 years ago

Hi, nlpnet is cool, but i cannot find good tutorial on how to train my custom Semantic Role labeling model (language other than english). I read docs http://nilc.icmc.usp.br/nlpnet/training-srl.html and tried to reproduce, but first problem i faced is that there is no such file nlpnet-train.py. I need to get new SRL model from text, so i tried to run in console: python train_srl.py srl pred --gold train_google.txt --data srl-model/

where train_google.txt - is txt file with raw text, train_srl.py is file in \Lib\site-packages\nlpnet\srl\ But nothing happens, no errors, no new folder created Am i right?

erickrf commented 8 years ago

Hi, you have to use nlpnet-train.py; train_srl.py doesn't train any models.

When you install nlpnet, nlpnet-train.py is copied from the bin/ directory to the Scripts folder of your Python installation or /usr/local/bin. If in Windows, are you sure it is in your PATH environment variable?

GraphGrailAi commented 8 years ago

Found it, close