Principle:Google deepmind Mujoco Model Text Dump
| Knowledge Sources | |
|---|---|
| Domains | Debugging, Model_IO |
| Last Updated | 2026-02-15 06:00 GMT |
Overview
Process of generating a human-readable text representation of all compiled model data for inspection and debugging.
Description
Model Text Dump produces a formatted text file containing every field of the compiled mjModel: sizes, options, body properties, joint parameters, geometry specifications, actuator settings, sensor configurations, and all numerical arrays. This is invaluable for debugging model compilation issues, verifying that the model was compiled as expected, and understanding the internal representation.
Usage
Use for debugging and inspection. The output file can be large for complex models (megabytes for humanoids). Compare text dumps of different model versions to identify changes.
Theoretical Basis
The dump iterates over all model fields in the X-macro-defined order, formatting each array with appropriate labels and precision. The output includes:
- Model size summary (number of bodies, joints, geoms, etc.)
- Options (timestep, solver settings, integrator)
- Per-element data (body masses, joint ranges, geom sizes)
- Full numerical arrays