See our collection for all versions of DeBERTa (v1 / v2 / v3).

Run DeBERTa with Keras 3: JAX, PyTorch, or TensorFlow

kerasformers/deberta_large

Papers: DeBERTa: Decoding-enhanced BERT with Disentangled Attention (arXiv:2006.03654) · DeBERTaV3 (arXiv:2111.09543) · HF Papers

DeBERTa is Microsoft's disentangled-attention text encoder (content + relative position). v1 uses byte-level BPE; v2/v3 use SentencePiece. v3 adds ELECTRA-style pretraining with gradient-disentangled embedding sharing. Import from deberta / deberta_v2 / deberta_v3 to match the generation.

For more details on the model, please go to the upstream model card.

Pure-Keras 3 conversion of microsoft/deberta-large for kerasformers. One implementation runs unmodified on TensorFlow / Torch / JAX.

This is a fill-mask / encoder checkpoint (DebertaMaskedLM, v1 large). Task heads (sequence/token classify, QA, …) load via hf: fine-tunes.

✨ Quick start (fill-mask)

import os
os.environ["KERAS_BACKEND"] = "torch"  # or "jax" / "tensorflow"

from kerasformers.models.deberta import (
    DebertaMaskedLM,
    DebertaTokenizer,
)

mlm = DebertaMaskedLM.from_weights("kerasformers/deberta_large")
tokenizer = DebertaTokenizer.from_weights("kerasformers/deberta_large")

inputs = tokenizer("The capital of France is [MASK].")
logits = mlm(inputs)  # (1, L, vocab_size)
mask = int((inputs["input_ids"][0] == tokenizer.mask_token_id).argmax())
print(tokenizer.decode([int(logits[0, mask].argmax())]))

Load any DeBERTa variant the same way with from_weights("kerasformers/<variant>"):

Variant	Hub	Generation
`deberta_base`	`kerasformers/deberta_base`	v1
`deberta_large`	`kerasformers/deberta_large`	v1
`deberta_v2_xlarge`	`kerasformers/deberta_v2_xlarge`	v2
`deberta_v2_xxlarge`	`kerasformers/deberta_v2_xxlarge`	v2
`deberta_v3_xsmall`	`kerasformers/deberta_v3_xsmall`	v3
`deberta_v3_small`	`kerasformers/deberta_v3_small`	v3
`deberta_v3_base`	`kerasformers/deberta_v3_base`	v3
`deberta_v3_large`	`kerasformers/deberta_v3_large`	v3

Tips

Set KERAS_BACKEND before importing Keras / kerasformers.
Prefer Tokenizer.from_weights(...) so vocab and mask token match.
Do not mix packages across generations (v1 ≠ v2 ≠ v3).
See DeBERTa docs and Loading Weights.
Community / upstream safetensors still work via the hf: prefix, e.g. DebertaMaskedLM.from_weights("hf:microsoft/deberta-large").