---
library_name: lucid
license: apache-2.0
tags:
  - masked-lm
  - bert
  - lucid
datasets:
  - wikipedia
  - bookcorpus
pipeline_tag: fill-mask
---

# BERT-Base (Masked-LM)

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

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

## Available weights

| Tag | Params | GFLOPs | Size | Source |
|---|---|---|---|---|
| `WIKIPEDIA_BOOKSCORPUS` *(default)* | 110.1M | — | 509.46 MB | transformers |

## Usage

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

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

# explicit tag (enum or string)
model = models.bert_base_mlm(weights=BertBaseMLMWeights.WIKIPEDIA_BOOKSCORPUS)
model = models.bert_base_mlm(pretrained="WIKIPEDIA_BOOKSCORPUS")

# 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  # (B, T, vocab_size)
```

## Conversion

Converted from `transformers/google-bert/bert-base-uncased` via
`python -m tools.convert_weights bert_base_mlm --tag WIKIPEDIA_BOOKSCORPUS`.
Key mapping + numerical parity verified against the source.

## License

`apache-2.0` — 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.
```