explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2025-07-17 19:52:18 +03:00
Code Issues Packages Projects Releases Wiki Activity

Update conll_train script

Browse Source
This commit is contained in:
Matthew Honnibal 2016-08-29 14:24:30 +02:00
parent 136a7a2322
commit 0c7520dbb7
1 changed files with 20 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
bin/parser/conll_train.py
View File

@ -119,7 +119,8 @@ def score_sents(nlp, gold_tuples):
def train(Language, gold_tuples, model_dir, dev_loc, n_iter=15, feat_set=u'basic', def train(Language, gold_tuples, model_dir, dev_loc, n_iter=15, feat_set=u'basic',
learn_rate=0.001, noise=0.01, update_step='sgd_cm', width=128, depth=3,
learn_rate=0.001, noise=0.01, update_step='sgd_cm', regularization=0.0,
batch_norm=False, seed=0, gold_preproc=False, force_gold=False): batch_norm=False, seed=0, gold_preproc=False, force_gold=False):
dep_model_dir = path.join(model_dir, 'deps') dep_model_dir = path.join(model_dir, 'deps')
pos_model_dir = path.join(model_dir, 'pos') pos_model_dir = path.join(model_dir, 'pos')
@ -132,11 +133,11 @@ def train(Language, gold_tuples, model_dir, dev_loc, n_iter=15, feat_set=u'basic
if feat_set != 'neural': if feat_set != 'neural':
Config.write(dep_model_dir, 'config', feat_set=feat_set, seed=seed, Config.write(dep_model_dir, 'config', feat_set=feat_set, seed=seed,
labels=ArcEager.get_labels(gold_tuples)) labels=ArcEager.get_labels(gold_tuples),
eta=learn_rate, rho=regularization)
else: else:
hidden_layers = [128] * 3 hidden_layers = [width] * depth
rho = 1e-4
Config.write(dep_model_dir, 'config', Config.write(dep_model_dir, 'config',
model='neural', model='neural',
seed=seed, seed=seed,
@ -148,18 +149,18 @@ def train(Language, gold_tuples, model_dir, dev_loc, n_iter=15, feat_set=u'basic
eta=learn_rate, eta=learn_rate,
mu=0.9, mu=0.9,
noise=noise, noise=noise,
ensemble_size=1, rho=regularization)
rho=rho)
nlp = Language(data_dir=model_dir, tagger=False, parser=False, entity=False) nlp = Language(data_dir=model_dir, tagger=False, parser=False, entity=False)
nlp.tagger = Tagger.blank(nlp.vocab, Tagger.default_templates()) nlp.tagger = Tagger.blank(nlp.vocab, Tagger.default_templates())
nlp.parser = BeamParser.from_dir(dep_model_dir, nlp.vocab.strings, ArcEager) #nlp.parser = BeamParser.from_dir(dep_model_dir, nlp.vocab.strings, ArcEager)
nlp.parser = Parser.from_dir(dep_model_dir, nlp.vocab.strings, ArcEager)
for word in nlp.vocab: for word in nlp.vocab:
word.norm = word.orth word.norm = word.orth
print(nlp.parser.model.widths) print(nlp.parser.model.widths)
print("Itn.\tP.Loss\tPruned\tTrain\tDev\tSize") print("Itn.\tP.Loss\tTrain\tDev\tnr_weight")
last_score = 0.0 last_score = 0.0
nr_trimmed = 0 nr_trimmed = 0
eg_seen = 0 eg_seen = 0
@ -197,9 +198,9 @@ def _train_epoch(nlp, gold_tuples, eg_seen, itn, dev_loc, micro_eval):
else: else:
dev_uas = 0.0 dev_uas = 0.0
train_uas = score_sents(nlp, micro_eval).uas train_uas = score_sents(nlp, micro_eval).uas
size = nlp.parser.model.mem.size size = nlp.parser.model.nr_weight
nr_upd = nlp.parser.model.time nr_upd = nlp.parser.model.time
print('%d,%d:\t%d\t%.3f\t%.3f\t%.3f\t%d' % (itn, nr_upd, int(loss), nr_trimmed, print('%d,%d:\t%d\t%.3f\t%.3f\t%d' % (itn, nr_upd, int(loss),
train_uas, dev_uas, size)) train_uas, dev_uas, size))
loss = 0 loss = 0
return eg_seen return eg_seen
@ -213,20 +214,26 @@ def _train_epoch(nlp, gold_tuples, eg_seen, itn, dev_loc, micro_eval):
batch_norm=("Use batch normalization and residual connections", "flag", "b"), batch_norm=("Use batch normalization and residual connections", "flag", "b"),
update_step=("Update step", "option", "u", str), update_step=("Update step", "option", "u", str),
learn_rate=("Learn rate", "option", "e", float), learn_rate=("Learn rate", "option", "e", float),
gradient_noise=("Gradient noise", "option", "w", float), regularization=("Regularization penalty", "option", "r", float),
neural=("Use neural network?", "flag", "N") gradient_noise=("Gradient noise", "option", "W", float),
neural=("Use neural network?", "flag", "N"),
width=("Width of hidden layers", "option", "w", int),
depth=("Number of hidden layers", "option", "d", int),
) )
def main(train_loc, dev_loc, model_dir, n_iter=15, neural=False, batch_norm=False, def main(train_loc, dev_loc, model_dir, n_iter=15, neural=False, batch_norm=False,
learn_rate=0.001, gradient_noise=0.0, update_step='sgd_cm'): width=128, depth=3, learn_rate=0.001, gradient_noise=0.0, regularization=0.0,
update_step='sgd_cm'):
with io.open(train_loc, 'r', encoding='utf8') as file_: with io.open(train_loc, 'r', encoding='utf8') as file_:
train_sents = list(read_conll(file_)) train_sents = list(read_conll(file_))
# Preprocess training data here before ArcEager.get_labels() is called # Preprocess training data here before ArcEager.get_labels() is called
train_sents = PseudoProjectivity.preprocess_training_data(train_sents) train_sents = PseudoProjectivity.preprocess_training_data(train_sents)
nlp = train(English, train_sents, model_dir, dev_loc, n_iter=n_iter, nlp = train(English, train_sents, model_dir, dev_loc, n_iter=n_iter,
width=width, depth=depth,
feat_set='neural' if neural else 'basic', feat_set='neural' if neural else 'basic',
batch_norm=batch_norm, batch_norm=batch_norm,
learn_rate=learn_rate, learn_rate=learn_rate,
regularization=regularization,
update_step=update_step, update_step=update_step,
noise=gradient_noise) noise=gradient_noise)
@ -237,6 +244,5 @@ def main(train_loc, dev_loc, model_dir, n_iter=15, neural=False, batch_norm=Fals
print('LAS', scorer.las) print('LAS', scorer.las)
if __name__ == '__main__': if __name__ == '__main__':
plac.call(main) plac.call(main)