Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Showing content from https://flairnlp.github.io/docs/tutorial-basics/tagging-entities below:

Tagging entities | flair

This tutorials shows you how to do named entity recognition, showcases various NER models, and provides a full list of all NER models in Flair.

Our standard model uses Flair embeddings and was trained over the English CoNLL-03 task and can recognize 4 different entity types. It offers a good tradeoff between accuracy and speed.

As example, let's use the sentence "George Washington went to Washington.":

from flair.nn import Classifier
from flair.data import Sentence


tagger = Classifier.load('ner')


sentence = Sentence('George Washington went to Washington.')


tagger.predict(sentence)


print(sentence)

This should print:

Sentence: "George Washington went to Washington ." → ["George Washington"/PER, "Washington"/LOC]

The printout tells us that two entities are labeled in this sentence: "George Washington" as PER (person) and "Washington" as LOC (location).

Our best 4-class model is trained using a very large transformer. Use it if accuracy is the most important to you, and speed/memory not so much.

from flair.data import Sentence
from flair.nn import Classifier


sentence = Sentence('George Washington went to Washington.')


tagger = Classifier.load('ner-large')


tagger.predict(sentence)


print(sentence)

As you can see, it's the same code, just with 'ner-large' as model instead of 'ner'. This model also works with most languages.

:::hint If you want the fastest model we ship, you can also try 'ner-fast'. :::

We also have NER models for text in other languages.

To tag a German sentence, just load the appropriate model:

tagger = Classifier.load('de-ner-large')


sentence = Sentence('George Washington ging nach Washington.')


tagger.predict(sentence)


print(sentence)

This should print:

Sentence: "George Washington ging nach Washington ." → ["George Washington"/PER, "Washington"/LOC]

Flair also works for languages that write from right to left. To tag an Arabic sentence, just load the appropriate model:

tagger = Classifier.load('ar-ner')


sentence = Sentence("احب برلين")


tagger.predict(sentence)


print(sentence)

This should print:

Sentence[2]: "احب برلين" → ["برلين"/LOC]

We also ship models that distinguish between more than just 4 classes. For instance, use our ontonotes models to classify 18 different types of entities.

from flair.data import Sentence
from flair.nn import Classifier


sentence = Sentence('On September 1st George won 1 dollar while watching Game of Thrones.')


tagger = Classifier.load('ner-ontonotes-large')


tagger.predict(sentence)


print(sentence)

This should print:

Sentence[13]: "On September 1st George won 1 dollar while watching Game of Thrones." → ["September 1st"/DATE, "George"/PERSON, "1 dollar"/MONEY, "Game of Thrones"/WORK_OF_ART]

Finding for instance that "Game of Thrones" is a work of art and that "September 1st" is a date.

For biomedical data, we offer the hunflair models that detect 5 different types of biomedical entities.

from flair.data import Sentence
from flair.nn import Classifier


sentence = Sentence('Behavioral abnormalities in the Fmr1 KO2 Mouse Model of Fragile X Syndrome.')


tagger = Classifier.load('bioner')


tagger.predict(sentence)


print(sentence)

This should print:

Sentence[13]: "Behavioral abnormalities in the Fmr1 KO2 Mouse Model of Fragile X Syndrome." → ["Behavioral abnormalities"/Disease, "Fmr1"/Gene, "Mouse"/Species, "Fragile X Syndrome"/Disease]

Thus finding entities of classes "Species", "Disease" and "Gene" in this text.

We end this section with a list of all models we currently ship with Flair.

You choose which pre-trained model you load by passing the appropriate string to the load() method of the Classifier class.

A full list of our current and community-contributed models can be browsed on the model hub.

RetroSearch is an open source project built by @garambo | Open a GitHub Issue

Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo

HTML: 3.2 | Encoding: UTF-8 | Version: 0.7.4