Datasets:
2 days ago
-
Add contact NRD architecture ablation (1M/1T/3M/3T): open-loop results and checkpoints (#2)
- Add contact NRD architecture ablation (1M/1T/3M/3T): open-loop results and checkpoints (b6506b04d0a590b89cc97e2004e6c3134d0cfea0)
- Remove videos (2aee928f0b38432c10334f2f83c6e4fca7ff7c87)
- Remove fair_1t (renamed) (4277e5c683003d32622359cabfb1ff5d41d1a027)
- Rename fair_1t to 1t_3t_recipe, update README (682c035a915905cbc2f4924c44a221b064351f5d)
-
g1_footstep_rl: release manifests of the G1 contact NRD case and the footstep policy (bytes, sha256 and provenance of every file)
-
g1_footstep_rl: G1 contact NRD (data, test, checkpoints) and the footstep policy (policies, RL inputs, MuJoCo results)
-
contact_nrd: correct the test-set provenance in the release manifest
-
contact_nrd: release manifest (sha256 of every file)
-
contact_nrd: training data, test sets and checkpoints for ball, pool and SO101 push-T