One-step action heads speed up vision-language-action models but reduce success
Technical Report: One-Step Drifting Action Heads for GR00T N1.7
Robotics
Summary
Generating robot actions in one step instead of many can make these systems faster, but it may cause the robot to fail more often. The authors tested a faster action method on a vision-language-action model called GR00T N1.7. They found the new method made the system run several times quicker but led to fewer successful task completions. This shows a trade-off between speed and accuracy in robot action planning.
What this means in practice
- •For robotics engineers: Optimize robot action modules when fast inference is critical but some decrease in task success is acceptable.
- •For real-time system developers: Design faster vision-language-action systems that manage trade-offs between speed and accuracy for timely responses.
Authors
Xihe Shao
Abstract
One-step action generation can substantially reduce the inference cost of vision-language-action (VLA) policies, but its effect on closed-loop task success remains an open question. This technical report studies a GR00T N1.7 variant in which the iterative diffusion-transformer action head is replaced by a one-step drifting action head, together with an overlap-conditioned extension for asynchronous chunk replacement. All multi-seed drifting runs were trained on two NVIDIA A800 GPUs. On LIBERO, the action head reduces the mean model-forward time of the action head from approximately $45.3\,\mathrm{ms}$ to $5.0\,\mathrm{ms}$, while the measured backbone-plus-head time falls from approximately $70.0\,\mathrm{ms}$ to $30.6\,\mathrm{ms}$. However, this speedup is accompanied by a systematic reduction in task success. Across three drifting seeds, success is $64.0\pm4.0\%$ on LIBERO-Spatial, $52.0\pm1.0\%$ on LIBERO-Goal, and $26.0\pm2.6\%$ on LIBERO-Long. The low seed variance indicates that the degradation is not explained by random initialization alone. We report the result as a speed--success trade-off rather than an overall improvement, and discuss likely contributing factors including deterministic one-step mode averaging, batch-dependent geometry estimation, long open-loop chunk execution, and the fact that synchronous LIBERO evaluation does not exercise the asynchronous overlap path.