Remove contaminated (unseen-in-training) in-domain scenes replaced in the previous commit

Fix in-domain scene contamination: replace unseen scenes with verified training-set members (250Hz: ep_000055/110/385 -> 056/109/383; 25Hz: ep_000788 -> 787). Membership now enforced by make_eval_scenes.py --train_camera_jsonl.

Ship frozen OOD scenes; seed sampling is not portable

Add ood_scenes_s20_250 so OOD eval replays the exact scenes instead of re-deriving from seeds

Document the bundled in-domain eval scenes

Add indomain_scenes_s20_250 so in-domain eval works without the source H5s

Document the raw event stream format and how to read it

Add raw v2e event streams (xytp + per-frame index), 1016 episodes

Remove raw event H5s uploaded to the repo root by mistake

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Upload README.md with huggingface_hub

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

initial commit