support for rerank model

[joebnb]

as title,current llama.cpp was supported rerank model and llama-server supported API, would lm studio support this?

ggml-org/llama.cpp#9510

[kalle07]

other embedder model support would be nice ...

RichardErkhov/BAAI_-_bge-reranker-v2-gemma-gguf
mradermacher/gte-Qwen2-7B-instruct-GGUF
second-state/jina-embeddings-v2-base-de-GGUF

HIT_TMG (he has checked all and is only a bit different llama implementation)
https://huggingface.co/HIT-TMG/KaLM-embedding-multilingual-mini-instruct-v1.5-GGUF
https://huggingface.co/HIT-TMG/KaLM-embedding-multilingual-mini-instruct-v1.5/discussions/8

[mNandhu]

+1 Support for Qwen3 Rerankers too?

[meotisR]

POST /v1/rerank when

[ciekawy]

There is now also qwen3 embedding/reranker set

…

On Sat, 21 Jun 2025 at 02:22, Shakel Roman ***@***.***> wrote: *meotisR* left a comment (lmstudio-ai/lmstudio-bug-tracker#470) <#470 (comment)> POST /v1/rerank when — Reply to this email directly, view it on GitHub <#470 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAVXJUDIICOG2ZMF3X5WKVL3ESQTFAVCNFSM6AAAAABYBY4O2KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDSOJTGE4DOMBQG4> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[iEddie-cmd]

Is ist implemented yet?

[kalle07]

atm all emedder models are working ... if not in list you only must setup manual override in model settings

but if its really works like an rerank-model in dont know ?!?

btw qwen is a bad embedder and that big in size and slow

[iEddie-cmd]

Its currently one of the top https://huggingface.co/spaces/mteb/leaderboard

[kalle07]

not with my setup and iv tested 20 embedders ...
maybe after 3 years of embedder they optimise it for the leaderboard-test ? ;)

[iEddie-cmd]

Which one do you use? i ll give it a try

[kalle07]

if you ask like this ... depends oc on content and setup and providing your docs

https://huggingface.co/kalle07/embedder_collection