RekaReka Inverse Dynamics
RIDM Pixel
The pixel model predicts camera movement directly from video frames. Its 9,836,063 parameters produce one prediction per frame.
Research release
What the release provides
This preliminary Apache 2.0 checkpoint was trained on game renders. The repository presents it as a candidate for further work on game footage.Reka Inverse Dynamics Model repository
How to use it well
Practical guidance. Preserve frame timing when preparing inputs. Compare motion labels with known camera controls before using them to train another model.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |