| license: mit | |
| library_name: keras | |
| pipeline_tag: feature-extraction | |
| tags: | |
| - keras | |
| - kerasformers | |
| - tf | |
| - jax | |
| - pytorch | |
| - backend-agnostic | |
| - deberta_v3 | |
| - feature-extraction | |
| base_model: | |
| - microsoft/deberta-v3-base | |
| # DeBERTa-v3 (deberta_v3_base) | |
| A pure [Keras 3](https://keras.io) port of **DeBERTa-v3**, converted from [`microsoft/deberta-v3-base`](https://huggingface.co/microsoft/deberta-v3-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.deberta_v3 import DebertaV3Model | |
| model = DebertaV3Model.from_weights("kerasformers/deberta_v3_base") | |
| ``` | |