Add Gemma embeddings

[sonam-pankaj95]

New embedding model from @google

[sonam-pankaj95]

I tried adding #gemma3 in #embedanything🦀, although it's working, the embeddings generated are the same as sentence transformers. That's why we haven't merged it yet.

My intuition is that, sentence-transformers are doing post-processing over the embeddings generated.

https://github.com/StarlightSearch/EmbedAnything/tree/gemma3