LMQLRobust language for LLM prompting
LMQL enables advanced prompting for LLMs through a modularized approach. Its features include:
- Types and Templates: Define structured queries to ensure better output.
- Constraints: Control generation behavior and size with specific requirements.
- Optimizing Runtime: Efficient execution and operation across multiple backends.
- Nested Queries: Support for procedural programming in prompting, allowing for complex query compositions.
Use cases range from simple question answers to detailed generation tasks. With LMQL, switch between models effortlessly, leveraging the power of multiple LLM backends.
