Repository navigation
v0.1.7
Release Notes for LLM Interface v0.1.7
Overview
This minor release improves the resilience of the LLM interface and adds better model detection mechanisms.
Features
- Structured Output Capability: Improved the matching of structured output capabilities of models based on their names and versions.
- Thinking Process Integration: Introduced a
requires_thinkingattribute to theLLMInterfaceclass, allowing models like Deepseek-R1 to document their 'thinking' process which is required when using Ollama with structured outputs.
Improvements
- Error Handling: Error handling in the
LLMInterfaceclass when response caching fails. - Test Suite Reliability: Regression tests for structured output generation forces
llm_clientto use structured outputs.