Discussions

I would love to see an offering of a compressor model like microsoft/llmlingua-2, that we could use both for prompts and RAG results

I'm running on an internal benchmark and Voyage has been amazing, about 5% better than OpenAI Ada v3. I was just wondering, has the code model also been instruction fine-tuned? I'm finding that if I add the document flag the overall quality is equal or worse.

I've been experimenting with using VoyageAI embeddings for a project where we are using cosine similarity as a first step in matching semantic equivalence of documents. I've noticed that compared to other embedding models I've tried like OpenAI and Bedrock, the embeddings and hence cosine similarities generated by VoyageAI embeddings are on a much more compressed range. As an example, the docs in the Quick start tutorial example <https://docs.voyageai.com/docs/quickstart-tutorial> have very similar cosines even though the docs are all quite different. Not sure if I'm doing something wrong, but I ran the reranker code for that example too, and the reranked relevance scores match what are shown on that page. The cosines I get for this query and documents are shown below. ```python query = "When is Apple's conference call scheduled?" documents = [ "The Mediterranean diet emphasizes fish, olive oil, and vegetables, believed to reduce chronic diseases.", "Photosynthesis in plants converts light energy into glucose and produces essential oxygen.", "20th-century innovations, from radios to smartphones, centered on electronic advancements.", "Rivers provide water, irrigation, and habitat for aquatic species, vital for ecosystems.", "Apple’s conference call to discuss fourth fiscal quarter results and business updates is scheduled for Thursday, November 2, 2023 at 2:00 p.m. PT / 5:00 p.m. ET.", "Shakespeare's works, like 'Hamlet' and 'A Midsummer Night's Dream,' endure in literature." ] ``` > VoyageAI voyage-2 ```python array([0.57205128, 0.5865394 , 0.62985496, 0.56841758, 0.84377816, 0.56752833]) ``` > OpenAI text-embedding-3-small ```python array([-0.00529196, 0.02914636, 0.14654271, -0.02232341, 0.78637504, -0.00315503]) ``` Obviously they are all relative but it feels weird. Is this just the nature of the VoyageAI embeddings or am I possibly doing something wrong?

Hello I am doing a project for school and am trying to compare model sizes based on parameters. Would you be able to tell me the amount of parameters this model uses?

OpenAI's new embedding models seem to work pretty well across a number of european languages (French, Spanish, Italian etc.). I am thinking of switching from OpenAI to Voyage for embeddings. Have your models been trained across text data in a number of languages? If so, do you have any performance benchmarks for say French vs English etc? Thanks!

I would like to know the list of languages supported by the embedding models offered by Voyage AI? Especially does any of the model support Mongolian language? Both Cyrillic and Latin versions of the language?

Looking for any example/notebook with best practices to vectorize and embed an entire repository using Voyage SDK to evaluate the code embedding API.

Would it be possible to add either a header or to the body a field to indicate the rate limit balance? I know you suggest adding a delay between calls, but that results in needless delays when you have a small number of requests, and my attempts to track the balance myself tend to be out of sync with your system.

Hi there From the documentation, I understand that voyage uses llama2 tokenizer. Is this also true for the `voyage-code-2` model? Also, which specific llama2 tokenizer used - is it `meta-llama/Llama-2-70b-hf`? Asking since I want to use a tokenizer in `RecursiveCharacterTextSplitter.from_huggingface_tokenizer` before passing the chunked texts to the API for embedding. Thanks

I used Langchain as a wrapper to access Voyage AI Embedding. I tried to use `voyage-large-02` in my local notebook and there was no any issues. But when I deployed it to my CI/CD pipeline, which uses Kubernetes pod as the instance, there's an error said: `Retrying langchain_community.embeddings.voyageai.embed_with_retry.<locals>._embed_with_retry in 10.0 seconds as it raised RuntimeError: Voyage API Error. Message: {"detail": "Model voyage-large-02 is not supported. Supported models are ['voyage-code-2', 'voyage-02', 'voyage-2', 'voyage-01', 'voyage-lite-01', 'voyage-lite-01-instruct', 'voyage-lite-02-instruct']."}.` What was happen? is that a Python version issue or anything else?