GGUF version
Could we please have access to a GGUF version of Jina V4 to be used with Ollama ?
V3 never got converted 😞
SSS
Hi @sqliteai , my colleague is working on it along with some other tasks, so I’m not sure exactly when it will be done. Probably this week or next.
any update ?
+1 for GGUF version
sorry may I ask why jina-embedding-v4-gguf cut visual model out ?
is that gguf format not support ?
I have interest in jina-embedding-v4 , since it state that "multimodal and multilingual retrieval" support, but now gguf format seems not support ?
image will be back, we just need more test there. image is treated a bit different in gguf.
also we aim to figure out the smallest minimum working unit so that people can deploy it at scale without any redundancy.
thanks, I really no idea on the format issue.
since using LM Studio for gguf format .... thanks for work and I think I will wait you the full function ready before test
Thank you !
Tried it a bit on multilingual documents, mostly French.
Sadly, BGE-M3 is still giving better results for me.
But Jina V4 gave good results. Just not as much complete and accurate than BGE-M3.