RoboAtlas
EN
Models

Benchmark

GR00T N1.7 inference timing

Official site

PyTorch Eager

E2E latency

GR00T · GR00T N1.7

85.8 ms

Publisher evaluation · Reported; not independently reproduced

TensorRT Full Pipeline

E2E latency

GR00T · GR00T N1.7

27.9 ms

Publisher evaluation · Reported; not independently reproduced

torch.compile

E2E latency

GR00T · GR00T N1.7

48.6 ms

Publisher evaluation · Reported; not independently reproduced

Submit a source or correction