Thrilled to see VFlash gaining momentum! 🚀
It's amazing to see the excitement around fast visual planning for Physical AI. Huge thanks to @DrJimFan for his enthusiasm and support!
Looking forward to pushing the boundaries of System-1 VLMs and embodied intelligence together!
It's amazing to see the excitement around fast visual planning for Physical AI. Huge thanks to @DrJimFan for his enthusiasm and support!
Looking forward to pushing the boundaries of System-1 VLMs and embodied intelligence together!
The era of System One Vision Model is here! Introducing VFlash: parsing the visual world in parallel at unmatched speeds for Physical AI.View quoted post on X
Most embodied prediction tasks like object localization, robot trajectory planning, and dense 3D motion are fixed-length float vectors. For example, a bounding box is always 6 tokens, a 10-waypoint robot arm trajectory is 54: we know these beforehand. Instead of decoding one token at a time, we can output the entire geometric shape in one go.
VFlash enables up to 12x faster action decoding and a 5.1x overall robot inference speedup on edge GPUs like the Jetson Thor and DGX Spark. What’s surprising is that it ends up even more accurate than the VLM backbone, with up to 61% lower 3D trajectory error. The intuition is that VFlash avoids the cascading errors of autoregressive decoding, where every coordinate is written in ink and every later point loses the chance to backspace.
Everything that moves will be autonomous, and everything autonomous will need to think on its own chip. The next inference revolution is physical reasoning at the speed of reflex.
And as always, it’s open paper, open source! Link in thread: