GitHub - gozdesahin/Subword_Semantic_Role_Labeling: Semantic role labeling with subwords (character, character-ngram and morphology)

Semantic Role Labeling with subword composition

We provide sample training/testing scripts for different subword units under example_scripts folder.

Train SRL models on CoNLL-09 SRL training sets and test/evaluate trained models on CoNLL-09 evaluation sets for all languages.

simple_UNIT.sh: Trains/tests base SRL models for the given subword UNIT. Please check train.py for parameter descriptions.
ensemble_UNIT1_UNIT2_UNITn: Voting ensemble for the provided pretrained base SRL models (UNIT1, UNIT2, ..., UNITn).
sg_UNIT1_UNIT2_UNITn: Trains/tests a stack generalizer model from the predictions of pretrained base SRL models (UNIT1, UNIT2, ..., UNITn).

Language : Python 2.7
CUDA : Cuda compilation tools, release 8.0, V8.0.44
Libraries: PyTorch 0.2.0 post 3, numpy 1.13.0

Name		Name	Last commit message	Last commit date
Latest commit History 21 Commits
IO		IO
SUB		SUB
example_scripts		example_scripts
.gitignore		.gitignore
README.md		README.md
SGModel.py		SGModel.py
constants.py		constants.py
dataset.py		dataset.py
ensemble.py		ensemble.py
eval09.pl		eval09.pl
loader.py		loader.py
model.py		model.py
optim.py		optim.py
prewordvectors.py		prewordvectors.py
scorer.py		scorer.py
test.py		test.py
test_stack_gen.py		test_stack_gen.py
train.py		train.py
train_stack_gen.py		train_stack_gen.py