---
license: mit
library_name: keras
pipeline_tag: fill-mask
tags:
  - keras
  - kerasformers
  - tf
  - jax
  - pytorch
  - backend-agnostic
  - roberta
  - fill-mask
base_model:
  - FacebookAI/roberta-base
---
# RoBERTa (roberta_base)

A pure [Keras 3](https://keras.io) port of **RoBERTa**, converted from [`FacebookAI/roberta-base`](https://huggingface.co/FacebookAI/roberta-base) (Hugging Face Transformers).

The weights are **backend-agnostic**: the same checkpoint loads and runs identically under the JAX, TensorFlow, or PyTorch Keras backend (`KERAS_BACKEND=jax|tensorflow|torch`), via [kerasformers](https://github.com/IMvision12/KerasFormers).

## Usage

```python
from kerasformers.models.roberta import RobertaModel

model = RobertaModel.from_weights("kerasformers/roberta_base")
```