Finetuning

#14

by HuggySSO - opened Jul 29

Discussion

HuggySSO

Jul 29

How well does this model perform on sts tasks in german and how would i finetune it on domain-specific data?

bobox

Jul 30

Uhm... I see that the base model is DeBERTa v3 (imo a clever choices, I also use that as base models), there should be also a DeBERTa v3 that is multilingual, but I don't know if mbxai plan to release a model based on that.

Other popular base model used as base for those embedders is xml-roberta that is multilingual

aamirshakir

Mixedbread org Aug 5

Hey @HuggySSO , please check our german model here. @bobox this here is a BERT model :). We used DeBERTa v3 for our reranking model.

aamirshakir changed discussion status to closed Aug 5

syed-adbul

Aug 23

Is it possible to fine-tune the model using let's say sentence-transformer on my own English data?

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment