lm-format-enforcer enforces the output format (e.g., JSON schema or regular
expressions) of language models. It supports popular LLM libraries and
integrates with frameworks like HuggingFace and LlamaCpp.