You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Qwen just published embedding and reranker models, performance is very good and these models support multi-language quite well. When could we support these models in openvino model server?
Also, both models support chat template and requires instructions while current embedding and rerank api doesn't support these inputs and only take raw strings. That means application should apply chat template by themselves. That's a quite tight coupling between application and the model, although we have a model server. Could model server take care of applying chat template instead?