text-embeddings-inference icon indicating copy to clipboard operation
text-embeddings-inference copied to clipboard

Support for deberta-v2 model variant

Open w3iw3i opened this issue 1 year ago • 6 comments

Feature request

When attempting to use this reranker model mxbai-rerank-large-v1 from huggingface with TEI, I got the following error message:

Error: Could not create backend

Caused by:
    Could not start backend: Model is not supported: unknown variant `deberta-v2`, expected one of `bert`, `xlm-roberta`, `camembert`, `roberta`, `distilbert`, `nomic_bert` at line 21 column 28

Steps to reproduce

model=mixedbread-ai/mxbai-rerank-large-v1
volume=$PWD/models

docker run -p 8088:80 -v $volume:/models --pull always ghcr.io/huggingface/text-embeddings-inference:cpu-1.1.0 --model-id $model

Motivation

Support for deberta-v2 model variant would allow the usage of mxbai-rerank-large-v1 model with /rerank endpoint

Your contribution

Support for deberta-v2 model variant

w3iw3i avatar Mar 13 '24 03:03 w3iw3i

Any updates?

sapountzis avatar Apr 22 '24 08:04 sapountzis

Yeah, it would be nice to get support for the DeBERTa architecture as well.

@OlivierDehaene is this model architecture on the roadmap? can we maybe get an idea of the models this library is planning on supporting with upcoming releases?

vikram71198 avatar May 07 '24 00:05 vikram71198

  • also looking for it if mxbai models can be added

touhi99 avatar May 21 '24 15:05 touhi99

please add deberta

Stealthwriter avatar May 30 '24 20:05 Stealthwriter

Are there any updates on this issue? it would be great if deberta is added

sufeng-niu avatar Jun 26 '24 22:06 sufeng-niu