vLLM Adds Day-One Support for Thinking Machines’ Inkling 2 MiniMachine translation
vLLM 首日支持 Thinking Machines 的 Inkling 2 Mini
vLLM supports Thinking Machines’ Inkling 2 Mini from day one. The looped mixture-of-experts model runs its middle layers three times; vLLM supports per-pass KV caches, fused kernels and DSpark speculative decoding.
This source permits summary display only.
Read at the original source