def test_train_charlm_changed_chache_load_use_tagger():
corpus = NLPTaskDataFetcher.fetch_data(NLPTask.FASHION)
tag_dictionary = corpus.make_tag_dictionary('ner')
# make a temporary cache directory that we remove afterwards
os.makedirs('./results/cache/', exist_ok=True)
embeddings = CharLMEmbeddings('news-forward-fast', cache_directory='./results/cache/')
tagger: SequenceTagger = SequenceTagger(hidden_size=256,
embeddings=embeddings,
tag_dictionary=tag_dictionary,
tag_type='ner',
use_crf=False)
# initialize trainer
trainer: SequenceTaggerTrainer = SequenceTaggerTrainer(tagger, corpus, test_mode=True)
trainer.train('./results', learning_rate=0.1, mini_batch_size=2, max_epochs=3)
# remove the cache directory
shutil.rmtree('./results/cache')
loaded_model: SequenceTagger = SequenceTagger.load_from_file('./results/final-model.pt')
sentence = Sentence('I love Berlin')
sentence_empty = Sentence(' ')
loaded_model.predict(sentence)
loaded_model.predict([sentence, sentence_empty])
loaded_model.predict([sentence_empty])
# clean up results directory
shutil.rmtree('./results')
def test_training():
corpus = NLPTaskDataFetcher.fetch_data(NLPTask.FASHION)
tag_dictionary = corpus.make_tag_dictionary('ner')
embeddings = WordEmbeddings('glove')
tagger: SequenceTagger = SequenceTagger(hidden_size=256,
embeddings=embeddings,
tag_dictionary=tag_dictionary,
tag_type='ner',
use_crf=False)
# initialize trainer
trainer: SequenceTaggerTrainer = SequenceTaggerTrainer(tagger,
corpus,
test_mode=True)
trainer.train('./results',
learning_rate=0.1,
mini_batch_size=2,
max_epochs=10)
# clean up results directory
shutil.rmtree('./results')
def test_train_load_use_tagger():
corpus = NLPTaskDataFetcher.fetch_data(NLPTask.FASHION)
tag_dictionary = corpus.make_tag_dictionary('ner')
embeddings = WordEmbeddings('glove')
tagger: SequenceTagger = SequenceTagger(hidden_size=256,
embeddings=embeddings,
tag_dictionary=tag_dictionary,
tag_type='ner',
use_crf=False)
# initialize trainer
trainer: SequenceTaggerTrainer = SequenceTaggerTrainer(tagger, corpus, test_mode=True)
trainer.train('./results', learning_rate=0.1, mini_batch_size=2, max_epochs=3)
loaded_model: SequenceTagger = SequenceTagger.load_from_file('./results/final-model.pt')
sentence = Sentence('I love Berlin')
sentence_empty = Sentence(' ')
loaded_model.predict(sentence)
loaded_model.predict([sentence, sentence_empty])
loaded_model.predict([sentence_empty])
# clean up results directory
shutil.rmtree('./results')
def train(data_dir,
model_dir,
hidden_dim=256,
word_embeddings='de-fasttext',
use_crf=True,
learning_rate=.1,
batch_size=32,
max_epochs=50,
seed=0,
dev_size=.1,
test_size=.2):
tag_type = 'ner'
corpus: TaggedCorpus = load_corpus(data_dir,
dev_size,
test_size,
seed,
tag_to_biloes=None)
print('Corpus:', corpus)
tag_dictionary = corpus.make_tag_dictionary(tag_type=tag_type)
print('Size of tag dictionary:', len(tag_dictionary))
embedding_types: List[TokenEmbeddings] = [
WordEmbeddings(word_embeddings),
# comment in this line to use character embeddings
#CharacterEmbeddings(),
# comment in these lines to use contextual string embeddings
# CharLMEmbeddings('news-forward'),
# CharLMEmbeddings('news-backward'),
]
embeddings: StackedEmbeddings = StackedEmbeddings(
embeddings=embedding_types)
tagger: SequenceTagger = SequenceTagger(hidden_size=hidden_dim,
embeddings=embeddings,
tag_dictionary=tag_dictionary,
tag_type=tag_type,
use_crf=use_crf)
trainer: SequenceTaggerTrainer = SequenceTaggerTrainer(tagger,
corpus,
test_mode=False)
trainer.train(model_dir,
learning_rate=learning_rate,
mini_batch_size=batch_size,
max_epochs=max_epochs)
def test_train_charlm_nochache_load_use_tagger(results_base_path,
tasks_base_path):
corpus = NLPTaskDataFetcher.fetch_data(NLPTask.FASHION,
base_path=tasks_base_path)
tag_dictionary = corpus.make_tag_dictionary('ner')
embeddings = CharLMEmbeddings('news-forward-fast', use_cache=False)
tagger: SequenceTagger = SequenceTagger(hidden_size=256,
embeddings=embeddings,
tag_dictionary=tag_dictionary,
tag_type='ner',
use_crf=False)
# initialize trainer
trainer: SequenceTaggerTrainer = SequenceTaggerTrainer(tagger,
corpus,
test_mode=True)
trainer.train(str(results_base_path),
learning_rate=0.1,
mini_batch_size=2,
max_epochs=2)
loaded_model: SequenceTagger = SequenceTagger.load_from_file(
results_base_path / 'final-model.pt')
sentence = Sentence('I love Berlin')
sentence_empty = Sentence(' ')
loaded_model.predict(sentence)
loaded_model.predict([sentence, sentence_empty])
loaded_model.predict([sentence_empty])
# clean up results directory
shutil.rmtree(results_base_path)
from models import FORWARD_LM, BACKWARD_LM, GLOVE
from embeddings import KeyedWordEmbeddings
from ne_groups import GROUPS
from corpora import read_group
embedding_types: List[TokenEmbeddings] = [
KeyedWordEmbeddings(GLOVE),
CharLMEmbeddings(FORWARD_LM),
CharLMEmbeddings(BACKWARD_LM)
]
embeddings: StackedEmbeddings = StackedEmbeddings(embeddings=embedding_types)
for entities in GROUPS:
corpus: TaggedCorpus = read_group(entities)
tag_dictionary = corpus.make_tag_dictionary(tag_type='ner')
tagger: SequenceTagger = SequenceTagger(hidden_size=512,
embeddings=embeddings,
tag_dictionary=tag_dictionary,
tag_type='ner',
use_crf=True)
trainer: SequenceTaggerTrainer = SequenceTaggerTrainer(tagger,
corpus,
test_mode=True)
file_name = '-'.join(entities)
trainer.train(f'data/models/{file_name}',
learning_rate=0.05,
mini_batch_size=124,
max_epochs=40,
save_model=True)
# comment in these lines to use contextual string embeddings
# CharLMEmbeddings('news-forward'),
# CharLMEmbeddings('news-backward'),
]
embeddings = WordEmbeddings("tmp/glove.bin")
# 5. initialize sequence tagger
from flair.models import SequenceTagger
tagger: SequenceTagger = SequenceTagger(hidden_size=1024,
embeddings=embeddings,
tag_dictionary=tag_dictionary,
tag_type=tag_type,
use_crf=True)
# 6. initialize trainer
from flair.trainers import SequenceTaggerTrainer
trainer: SequenceTaggerTrainer = SequenceTaggerTrainer(tagger, corpus)
# 7. start training
trainer.train('resources/taggers/example-ner',
learning_rate=0.1,
mini_batch_size=8,
max_epochs=150)
# 8. plot training curves (optional)
from flair.visual.training_curves import Plotter
plotter = Plotter()
plotter.plot_training_curves('resources/taggers1/example-ner/loss.tsv')
plotter.plot_weights('resources/taggers1/example-ner/weights.txt')