A curated collection of Vision-Language-Action (VLA) models by SberRoboticsCenter, optimized in GGUF format for local inference.