William FH

I think most of the vector stores "scores" by default are distances, though it sometimes depends on how you configure the underlying index. This is why the relevance scores method...

It's a restriction on the Google API level, rather than the LangChain integration. I could see a use case where the bot leads in a conversation (system info gives background,...

Would happily review a PR of a notebook that reliably implements this with OSS models! My laptop crashes anytime I try to run anything > 1B parameters in size

You can likely replace the system messages with user messages

I think you'd need to `yield` the token outputs in the call_model function so that it is treated as a generator. @eyurtsev anything else required to get the streaming to...

Thanks for the PR, @adosaiguas ! Excited for your involvement. We plan to limit the scope of the `langgraph` package + repo to be core implementations + API's (similar to...

Thank you!

Ya we plan to change when we get the time. Would be happy to review a PR if anyone makes one here!

@russell-dot-js thanks for flagging. It typically _does_ work my exposing the runnable directly `graph.compile()` , but it looks like there may be some bug in how we're returning the response...

Possibly resolved in https://github.com/langchain-ai/langchain/issues/18741 and related to #136