README.md

---
library_name: transformers
tags:
- multilabel
- multilabel-token-classification
base_model:
- google-bert/bert-large-cased
---
# Overview
- This is an extension of the `bert-large-cased` model to enable **multi-label token classification**.
- The training objective is BCELoss.
- Labels are one-hot encoded.
- Model output logits can be normalized using sigmoid activation.
- This model uses the same weights as `bert-large-cased` and thus needs to be fine-tuned for downstream tasks.

# Usage
To initialize the model for fine tuning, simply provide `id2label` and `label2id`, similarly to standard token classification fine tuning:
```python
from transformers import AutoModelForTokenClassification

model = AutoModelForTokenClassification.from_pretrained('jvaquet/multilabel-classification-bert', 
  id2label = id2label, 
  label2id = label2id,
  trust_remote_code=True)
```