Description
ALBERT Model with a token classification head on top (a linear layer on top of the hidden-states output) e.g. for Named-Entity-Recognition (NER) tasks.
albert_base_token_classifier_conll03 is a fine-tuned ALBERT model that is ready to use for Named Entity Recognition and achieves state-of-the-art performance for the NER task. This model has been trained to recognize four types of entities: location (LOC), organizations (ORG), person (PER), and Miscellaneous (MISC).
We used TFAlbertForTokenClassification to train this model and used AlbertForTokenClassification annotator in Spark NLP 🚀 for prediction at scale!
Predicted Entities
LOC, ORG, PER, MISC
How to use
document_assembler = DocumentAssembler() \
.setInputCol('text') \
.setOutputCol('document')
tokenizer = Tokenizer() \
.setInputCols(['document']) \
.setOutputCol('token')
tokenClassifier = AlbertForTokenClassification \
.pretrained('albert_base_token_classifier_conll03', 'en') \
.setInputCols(['token', 'document']) \
.setOutputCol('ner') \
.setCaseSensitive(True) \
.setMaxSentenceLength(512)
# since output column is IOB/IOB2 style, NerConverter can extract entities
ner_converter = NerConverter() \
.setInputCols(['document', 'token', 'ner']) \
.setOutputCol('entities')
pipeline = Pipeline(stages=[
document_assembler,
tokenizer,
tokenClassifier,
ner_converter
])
example = spark.createDataFrame([['My name is John!']]).toDF("text")
result = pipeline.fit(example).transform(example)
val document_assembler = DocumentAssembler()
.setInputCol("text")
.setOutputCol("document")
val tokenizer = Tokenizer()
.setInputCols("document")
.setOutputCol("token")
val tokenClassifier = AlbertForTokenClassification.pretrained("albert_base_token_classifier_conll03", "en")
.setInputCols("document", "token")
.setOutputCol("ner")
.setCaseSensitive(true)
.setMaxSentenceLength(512)
// since output column is IOB/IOB2 style, NerConverter can extract entities
val ner_converter = NerConverter()
.setInputCols("document", "token", "ner")
.setOutputCol("entities")
val pipeline = new Pipeline().setStages(Array(document_assembler, tokenizer, tokenClassifier, ner_converter))
val example = Seq.empty["My name is John!"].toDS.toDF("text")
val result = pipeline.fit(example).transform(example)
import nlu
nlu.load("en.classify.token_albert_base_token_classifier_conll03").predict("""My name is John!""")
Results
+------------------------------------------------------------------------------------+
|result |
+------------------------------------------------------------------------------------+
|[B-PER, I-PER, O, O, O, B-LOC, O, O, O, B-LOC, O, O, O, O, B-PER, O, O, O, O, B-LOC]|
+------------------------------------------------------------------------------------+
Model Information
| Model Name: | albert_base_token_classifier_conll03 |
| Compatibility: | Spark NLP 3.3.0+ |
| License: | Open Source |
| Edition: | Official |
| Input Labels: | [token, document] |
| Output Labels: | [ner] |
| Language: | en |
| Case sensitive: | false |
| Max sentense length: | 512 |
Data Source
https://www.clips.uantwerpen.be/conll2003/ner/
Benchmarking
precision recall f1-score support
B-LOC 0.95 0.97 0.96 1837
B-MISC 0.87 0.86 0.87 922
B-ORG 0.90 0.91 0.90 1341
B-PER 0.91 0.97 0.94 1842
I-LOC 0.88 0.86 0.87 257
I-MISC 0.78 0.76 0.77 346
I-ORG 0.84 0.85 0.85 751
I-PER 0.97 0.92 0.94 1307
O 0.99 0.99 0.99 42759
accuracy 0.98 51362
macro avg 0.90 0.90 0.90 51362
weighted avg 0.98 0.98 0.98 51362
processed 51362 tokens with 5942 phrases; found: 6182 phrases; correct: 5382.
accuracy: 91.82%; (non-O)
accuracy: 98.01%; precision: 87.06%; recall: 90.58%; FB1: 88.78
LOC: precision: 93.76%; recall: 95.70%; FB1: 94.72 1875
MISC: precision: 80.55%; recall: 83.08%; FB1: 81.79 951
ORG: precision: 84.12%; recall: 87.32%; FB1: 85.69 1392
PER: precision: 85.90%; recall: 91.59%; FB1: 88.65 1964