has anyone had success using local hf models with rag evaluators (faithfulness, correctness, etc)? if so, which model(s) worked?

i know i can customize the prompt templates but im hoping to stay "out-of-box" so i can more easily stay up to date with upstream

Anyone having issues with Anthropic tool calling and streaming? The text response includes tool names so i know its getting tool names, but no tools are getting called (not returning tool call responses, me thinks)...

hey y'all, sorry if this is a simple question, is there a way to get a RouterQueryEngine behind a chat_engine? Ive tried going the ObjectIndex route, but no luck...