def benchmark_flair_mdl():
tagger = load_flair_ner_model()
start = time.time()
flair_sentences = []
for i, sentence in enumerate(sentences_tokens):
flair_sentence = Sentence()
for token_txt in sentence:
flair_sentence.add_token(Token(token_txt))
flair_sentences.append(flair_sentence)
tagger.predict(flair_sentences, verbose=True)
predictions = [[tok.tags['ner'].value for tok in fs]
for fs in flair_sentences]
print('Flair:')
print("Made predictions on {} sentences and {} tokens in {}s".format(
num_sentences, num_tokens,
time.time() - start))
assert len(predictions) == num_sentences
print(
classification_report(sentences_entities,
remove_miscs(predictions),
digits=4))
def test_flair_tagger(self):
# Download model beforehand
download_model('flair.ner',
DEFAULT_CACHE_DIR,
process_func=_unzip_process_func,
verbose=True)
print("Downloaded the flair model")
# Load the NER tagger using the DaNLP wrapper
flair_model = load_flair_ner_model()
# Using the flair POS tagger
sentence = Sentence(
'jeg hopper på en bil som er rød sammen med Jens-Peter E. Hansen')
flair_model.predict(sentence)
expected_string = "jeg hopper på en bil som er rød sammen med Jens-Peter <B-PER> E. <I-PER> Hansen <I-PER>"
self.assertEqual(sentence.to_tagged_string(), expected_string)
def benchmark_flair_mdl():
tagger = load_flair_ner_model()
start = time.time()
flair_sentences = []
for i, sentence in enumerate(sentences_tokens):
flair_sentence = Sentence()
for token_txt in sentence:
flair_sentence.add_token(Token(token_txt))
flair_sentences.append(flair_sentence)
tagger.predict(flair_sentences, verbose=True)
predictions = [[tok.get_tag('ner').value for tok in fs]
for fs in flair_sentences]
print('Flair:')
print_speed_performance(start, num_sentences, num_tokens)
assert len(predictions) == num_sentences
print(f1_report(sentences_entities, remove_miscs(predictions), bio=True))
def setup(self):
# A horrible hack to fix a check in flair.models.language_model for pytorch version that fails on cpu
if "+cpu" in torch.__version__:
torch.__version__ = torch.__version__.replace("+cpu", "")
self.model = dm.load_flair_ner_model()
import time
# models
from danlp.models import load_bert_ner_model, load_flair_ner_model
# dataset
from danlp.datasets import DDT
# utils
from flair.data import Sentence, Token
from copy import deepcopy
from memory_profiler import profile
# load models
bert = load_bert_ner_model()
flair = load_flair_ner_model()
# get data (splitted into a training set, a validation set, and a test set)
ddt = DDT()
train, valid, test = ddt.load_as_simple_ner(True)
# divide the observations and the targets of the testset into new variables
sentences, categories = test
@profile
def get_bert_predictions(sentences):
predictions = []
'''for sentence in sentences:
predictions.append(bert.predict(sentence)[1])'''
pred = bert.predict(sentences[0])