Collected sources and patterns will appear here. Add from search, explore, or the patterns library.
Tensor<NormalizedActions> -> Tensor<PhysicalActions>
Decode predicted model action tokens and denormalize them using target-dataset calibration parameters.
Problem it solves
Raw VLA model action outputs are normalized/tokenized and must be converted back to real physical metrics to control a specific robot platform.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.