---
license: mit
task_categories: [robotics]
tags: [ci-mse, offline-evaluation, so100, maniskill, lerobot, domain-randomization]
pretty_name: CI-MSE × SO-100 bowl task — v2 experiment artifacts
---
# CI-MSE v2 — raw artifacts for the SO-100 "put the {color} cube in the bowl" study

| folder | contents |
|---|---|
| `raw/<split>/traj.h5, traj.json` | merged ManiSkill trajectories (obs images, qpos, actions, env states, success, phase info) per source |
| `raw/<split>/gen_stats_total.json` | oracle attempts / saved / failure breakdown |
| `media/proof_<split>/` | final-frame grids (named cube inside bowl) and phase strips — visual verification of task success |
| `media/videos_<split>/` | rendered oracle episodes |
| `media/so100_bowl_*/random_frames.png` | decoded random frames from the built LeRobot datasets |
| `verify/<dataset>.json` | automated verification report per dataset (counts, units, tasks, decode, colour balance) |
| `eval_seeds.json` | 200 closed-loop eval scene seeds (never used for demos), max_episode_steps 450 |

Datasets: `Kavin60606/so100_bowl_train`, `so100_bowl_val_s0`, `so100_bowl_val_s1`, `so100_bowl_val_s2`, `so100_bowl_val_mix`.
Env + oracle + build code: `fd-studio/eval/sim_so100_v2/`. Notes: `DATA_v2.md`.