AI studio inference server requirements are defined by GPU VRAM to hold model weights, sufficient system RAM for da