ARDY is an autoregressive diffusion model designed for interactive motion generation, supporting online text prompting and flexible long-horizon kinematic constraints (root paths/waypoints, full-body ...

I've been running Qwen 3.5 9B since I have 16GB VRAM, but this machine also has 64GB DRAM, is running something bigger feasible with spillage onto RAM or is the extra latency and splitting across VRAM ...