Chen Qian

Thanks for putting up the detailed feature request! > MCP and A2A which become defacto industry standard for now. MCP is good and getting widely adopted, while I don't feel...

@okhat @TomeHirata Maybe we should include a configurable argument in `ChatAdapter` to skip JSON retry? Default to JSON retry is useful, but sometimes this hits me when debugging DSPy programs...

@hung-phan Thanks for reporting the issue! We allow calling async tool in sync context: https://dspy.ai/tutorials/async/#using-async-tools-in-synchronous-contexts, which is the recommended way to run optimization with async tools. Could you try it...

@enitrat Thanks for reporting the issue! The mock makes sense to me, but are you spotting it in real LM calls? Adapters are built on top of a series of...

@enitrat could you share a reproducible code? We have never noticed this even on small models like 1b/3b size. By far LMs follow line breaking rule pretty well, so I...

@enitrat thanks! to clarify - `[[ ## field_name ## ]]` is not the full boilerplate for formatting/parsing purposes, but `\n\n[[ ## field_name ## ]]\n\n` is. So if we cannot stably...

@BoluoZz Thanks for reporting the issue! The json repair behavior change makes sense to me, but could you share a reproducible code (including your module and LM)? We did see...

@BoluoZz Thanks for the detailed information! Let me do some testing and update the deps accordingly.

@Haoping-Xiao Thanks for reporting the issue, could you share the code to reproduce?

I don't fully understand the issue, computing BLEU in DSPy shouldn't be different from other cases. You can define any custom metric that takes in the train/dev/test data, and the...