2.11.0 Release notes
This release includes an upgrade to HTML-based widgets for improved rendering and the addition of a G-eval test in the library.
Enhancements
Upgrade to HTML-based widgets for improved rendering
We've upgraded ValidMind's rendering engine to a new HTML-based system, ensuring interactive components are preserved when notebooks are saved.
Key improvements:
- Interactive charts remain fully functional after saving and reopening notebooks.
- Progress bars persist their state across notebook sessions.
- More reliable rendering with improved state management.
To disable interactive figures, set VALIDMIND_INTERACTIVE_FIGURES=false (default: enabled).
GEval evaluation metrics support
ValidMind now supports GEval LLM evaluation metrics from DeepEval, allowing you to define custom criteria for evaluations that are logged directly in your documentation via ValidMind tests.
GEval enables "LLM-as-judge" metrics, where a language model scores outputs according to explicit, user-defined standards, eliminating dependence on string matching or reference responses.