add embedding model(s)

Please add an embeddings endpoint to enable embedding generation from any model!

Please do NOT limit embedding generation to specialized embedding models. Many providers do not recognize that all models can create embeddings! Yes, embedding models are much smaller and run much faster than full-size models when creating embeddings, but the different model types apply to different use cases.

My company would bring a significant amount of inference demand to venice if embeddings from full-size models became available.

yes please. this would allow us to cut 100% of openai depencencies on our production systems

this would be really cool

I would 100% switch from openai’s API if anyone else offered decent embeddings. Especially if the context window and message sizes were larger than openai’s!

another +1 to this. the current status quo makes me use models from big tech which i dont want to do!

commented as Anonymous before, this would be really niece to have

this would be great, so we do not have to use OpenAI or Gemini embedding