---

license: apache-2.0 library_name: span-marker tags:

• span-marker
• token-classification
• ner
• named-entity-recognition pipeline_tag: token-classification widget:
  • text: >- Amelia Earhart flew her single engine Lockheed Vega 5B across the Atlantic to Paris . example_title: Amelia Earhart model-index:
  • name: >- SpanMarker w. xlm-roberta-large on CoNLL++ with document-level context by Tom Aarsen results:
    • task: type: token-classification name: Named Entity Recognition dataset: type: conllpp name: CoNLL++ w. document context split: test revision: 3e6012875a688903477cca9bf1ba644e65480bd6 metrics:
      • type: f1 value: 0.9554 name: F1
      • type: precision value: 0.9600 name: Precision
      • type: recall value: 0.9509 name: Recall datasets:
  • conllpp
  • tomaarsen/conllpp language:
  • en metrics:
  • f1
  • recall
  • precision

---

SpanMarker for Named Entity Recognition

This is a SpanMarker model that can be used for Named Entity Recognition. In particular, this SpanMarker model uses xlm-roberta-large as the underlying encoder. See train.py for the training script. Note that this model was trained with document-level context, i.e. it will primarily perform well when provided with enough context. It is recommended to call model.predict with a 🤗 Dataset with tokens, document_id and sentence_id columns. See the documentation of the model.predict method for more information.

Usage

To use this model for inference, first install the span_marker library:

pip install span_marker

You can then run inference with this model like so:

from span_marker import SpanMarkerModel

# Download from the 🤗 Hub
model = SpanMarkerModel.from_pretrained("tomaarsen/span-marker-xlm-roberta-large-conllpp-doc-context")
# Run inference
entities = model.predict("Amelia Earhart flew her single engine Lockheed Vega 5B across the Atlantic to Paris.")

Limitations

Warning: This model works best when punctuation is separated from the prior words, so

# ✅
model.predict("He plays J. Robert Oppenheimer , an American theoretical physicist .")
# ❌
model.predict("He plays J. Robert Oppenheimer, an American theoretical physicist.")

# You can also supply a list of words directly: ✅
model.predict(["He", "plays", "J.", "Robert", "Oppenheimer", ",", "an", "American", "theoretical", "physicist", "."])

The same may be beneficial for some languages, such as splitting "l'ocean Atlantique" into "l' ocean Atlantique".

See the SpanMarker repository for documentation and additional information on this library.