sglang-serving

Serve LLMs with SGLang for structured generation and high-throughput inference. Use when launching an SGLang server, using RadixAttention, producing constrained JSON/regex output, or performing structured decoding.