From one model to seven — what it took to make TurboQuant model-portable

turboquant-vllm started on Molmo2 only. v1.3.0 validates seven model families — after rewriting five Triton kernels and teaching the cache about sliding window attention.