stanfordnlp/dspy 3.0.3

on GitHub

What's Changed

New Functionality

• Introduce rollout_id for bypassing LM cache in a namespaced way by @okhat in #8745
• Raise temperature when doing multiple rollouts (and warn otherwise) by @okhat in #8748

Optimizers

• Bump gepa version - Fix dependencies, Add progress bar, Ensure failing metrics do not continue running by @LakshyAAAgrawal in #8701
• Fix call to forward in dspy.GEPA tutorial by @LakshyAAAgrawal in #8719

Maintenance

• Replace ujson by orjson by @chenmoneygithub in #8655
• Change how forward is patched to avoid warning on explicit forward call by @chenmoneygithub in #8700
• Add type hints and fix doc string of SIMBA by @TomeHirata in #8707
• Remove deprecated LiteLLM caching from LM by @okhat in #8742
• Fix Predict config merging and document rollout_id usage by @okhat in #8746
• Lower max_tokens requirement for reasoning LMs by @okhat in #8750

Full Changelog: 3.0.2...3.0.3