Best way of using DINOv2 image similarity in Weavite

General
1

Hello Weaviate gurus. Long time no see :frowning: How boring when I-m not using Weaviate :wink:
I started investigating the wonderful world of multimodal similarity hoping to get some work from a nice historical archive.
I played with CLIP and it is nice but in my experience it works okaish for text to image recognition but somehow fails miserably when asking for an image to image similarity (perceptual+semantic).

After some investigation I think I will want to test the CLIPv2 model.

Not sure as to the best way to use that aside from a ‘bring your own vectors’ approach.

Is there a smarter way?

Thanks a lot

2

Ciao amico mio! È passato così tanto tempo dall’ultima volta che ci siamo sentiti!

I am pretty sure you will find this interesting, then:

and, of course, source code here: GitHub - antas-marcin/cliparena-page: Website comparing find similar, image and text search using AI CLIP models

Enjoy!

3

Ciao amico mio. Yes I misssed you too :slight_smile: Boring work for monhts :slight_smile:

As always you are super helpful (and friendly) :slight_smile:

Will investigate.

Hope all good with you and wow the ‘ask AI - internal RAG we now have on the Weaviate forum is soo damn useful. GREAT job.

Take care.

Bob

1 Like