LLM continuous batching host scheduler in Rust (191k tok/s on 1 vCPU)

The LLM continuous batching host scheduler written in Rust delivers 191k tokens per second on a single vCPU, eliminating manual steps in workflows. Start by pro