Yann Stoneman’s Post

Thanks VentureBeat, for covering my blog about Cohere's new multimodal Embed 3 model! Unlike other models that separate text and image data into distinct clusters, Embed 3 maintains equal performance across both types in a shared vector space. This means your searches return the most relevant results regardless of content type - no more text-biased or image-biased results. That also means that orgs can upgrade to multimodal without re-vectorizing their existing data. #genai #datascience #llms #nlp #cohere #multimodal #search https://lnkd.in/eki6tu6v