On my foundry stack we observe Text Embedding 3 Large in model catalog, but not as an option for text embeddings in Pipeline Builder.
Is this behavior expected? Or when can we expect to be able to leverage Text Embedding 3 Large in pipeline builder’s embed text transform?
Hey there! Right now embeddings 3 large outputs an embedded vector that is too large to be input to the ontology. As a result, we have blocked support for embedding 3 large until the ontology can support indexing on the larger embedded vector!
Hey David, how about text-embedding-3-small?
Text to Embeddings 3 small is supported @josh1