---
library_name: lucid
license: mit
tags:
  - token-classification
  - bert
  - lucid
datasets:
  - conll2003
pipeline_tag: token-classification
model-index:
  - name: bert-base-token-cls
    results:
      - task: { type: token-classification }
        dataset: { name: conll2003, type: conll2003 }
        metrics:
          - { type: f1, value: 91.3 }
---

# BERT-Base (CoNLL-2003 NER)

> https://arxiv.org/abs/1810.04805

[Lucid](https://github.com/ChanLumerico/lucid) port of `transformers/dslim/bert-base-NER`,
converted to Lucid-native safetensors.

## Available weights

| Tag | f1 | Params | GFLOPs | Size | Source |
|---|---|---|---|---|---|
| `CONLL2003` *(default)* | 91.3 | 108.3M | — | 413.22 MB | transformers |

## Usage

```python
import lucid
import lucid.models as models
from lucid.models.weights import BERTBaseNERWeights

# default tag
model = models.bert_base_token_cls(pretrained=True)

# explicit tag (enum or string)
model = models.bert_base_token_cls(weights=BERTBaseNERWeights.CONLL2003)
model = models.bert_base_token_cls(pretrained="CONLL2003")

# feed token ids (tokenize with the matching lucid.utils.tokenizer)
input_ids = lucid.tensor([[101, 7592, 2088, 102]], dtype=lucid.int64)
out = model(input_ids)
logits = out.logits  # classification logits
```

## Conversion

Converted from `transformers/dslim/bert-base-NER` via
`python -m tools.convert_weights bert_base_token_cls --tag CONLL2003`.
Key mapping + numerical parity verified against the source.

## License

`mit` — inherited from the original weights.

## Citation

```
Devlin et al., "BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding", NAACL 2019. Miniatures: Turc et al., "Well-Read Students Learn Better", 2019.
```