---
license: mit
pretty_name: Memory-Manip-Bench
tags:
  - robotics
  - imitation-learning
  - diffusion-policy
  - memory
  - pomdp
  - robosuite
  - mujoco
size_categories:
  - 100K<n<1M
task_categories:
  - robotics
---

# Memory-Manip-Bench

Memory-dependent tabletop manipulation tasks for a Franka Panda (robosuite 1.5 /
MuJoCo 3), introduced in **"Remember what you did?: Learning Behavioral Memories
for Partially Observable Object Manipulation"**
([paper](https://arxiv.org/abs/2606.21188) · [project page](https://robo-camp.github.io/)). Every task hides one piece of
information that the policy can only discover by interacting and must then
remember; forgetting it makes the episode fail.

## Tasks

<table align="center">
<tr>
<td align="center"><img src="https://huggingface.co/datasets/harrywang01/Memory-Manip-Bench/resolve/main/media/uncover_blocks.gif" width="200"><br><b>Uncover-Blocks</b></td>
<td align="center"><img src="https://huggingface.co/datasets/harrywang01/Memory-Manip-Bench/resolve/main/media/swap_block.gif" width="200"><br><b>Swap-Block</b></td>
<td align="center"><img src="https://huggingface.co/datasets/harrywang01/Memory-Manip-Bench/resolve/main/media/probe_insert.gif" width="200"><br><b>Probe-Insert</b></td>
<td align="center"><img src="https://huggingface.co/datasets/harrywang01/Memory-Manip-Bench/resolve/main/media/button_lightbulb.gif" width="200"><br><b>Button-Lightbulb</b></td>
</tr>
<tr>
<td align="center"><img src="https://huggingface.co/datasets/harrywang01/Memory-Manip-Bench/resolve/main/media/stack_lego.gif" width="200"><br><b>Stack-Lego</b></td>
<td align="center"><img src="https://huggingface.co/datasets/harrywang01/Memory-Manip-Bench/resolve/main/media/find_soda.gif" width="200"><br><b>Find-Soda</b></td>
<td align="center"><img src="https://huggingface.co/datasets/harrywang01/Memory-Manip-Bench/resolve/main/media/open_door.gif" width="200"><br><b>Open-Door</b></td>
</tr>
</table>

| file | task | env (`env_name`) | task camera | what must be remembered |
|---|---|---|---|---|
| `image_uncoverblock.hdf5` | **Uncover-Blocks** | `UncoverBlockVisionHard` | `agentview` | which covers were already lifted |
| `image_blockswap.hdf5` | **Swap-Block** | `FruitSwapEasy` | `peg_focus_view` | where each cube started |
| `image_insertpeg.hdf5` | **Probe-Insert** | `InsertPegVision` | `peg_focus_view` | which holes were already tried |
| `image_buttonlightbulb.hdf5` | **Button-Lightbulb** | `ButtonLightbulbVision` | `agentview` | which button toggles which bulb |
| `image_legostacking.hdf5` | **Stack-Lego** | `LegoStackingVision` | `peg_focus_view` | which colour is the solid-bottom brick |
| `image_findsoda.hdf5` | **Find-Soda** | `FindSodaVision` | `drawerview` | which drawers were already checked |
| `image_opendoors.hdf5` | **Open-Door** | `OpenDoorsVision` | `doorview` | which doors were already tried |

One robomimic-style hdf5 of teleoperated demonstrations per task (100 demos each).

Environments, teleoperation scripts and the CAMP training code are in the
[CAMP repository](https://robo-camp.github.io/) (`memory_manip_bench` package).

## Download

```bash
pip install -U "huggingface_hub[cli]"
huggingface-cli download harrywang01/Memory-Manip-Bench --repo-type dataset \
    --include image_uncoverblock.hdf5 --local-dir data/
```

## Data format

```
data/                              (attrs: env_args — env_name + env_kwargs to rebuild the env)
└── demo_<i>/                      (attrs: num_samples, model_file; Find-Soda also ep_meta)
    ├── obs/<task camera>_image    (T, 256, 256, 3) uint8
    ├── obs/robot0_eye_in_hand_image (T, 256, 256, 3) uint8
    ├── obs/robot0_eef_pos         (T, 3)  float
    ├── obs/robot0_eef_quat        (T, 4)  float
    ├── obs/robot0_gripper_qpos    (T, 2)  float
    ├── actions                    (T, 7)  delta OSC command
    ├── actions_abs                (T, 7)  absolute [goal_pos(3), goal_axis_angle(3), gripper(1)]
    ├── rewards, dones             (T,)
    └── states                     (T, D)  MuJoCo states
```

```python
import h5py, json
with h5py.File("data/image_uncoverblock.hdf5", "r") as f:
    env_args = json.loads(f["data"].attrs["env_args"])
    demo = f["data/demo_1"]
    imgs, acts = demo["obs/agentview_image"][:], demo["actions_abs"][:]
```

`model_file` (the demo's MuJoCo model) and `states` replay a demo in the environment. Find-Soda's
placement target is a model parameter, not simulator state, so its demos also carry
`ep_meta = {"target_xy", "soda_drawer_idx"}`, which `FindSodaVision.set_ep_meta` restores on replay.

## Changelog

- **2026-10-11, `image_findsoda.hdf5`:** each demo's `model_file` now places the target patch where
  the images show it (under the placed can), and `ep_meta` was added. Before, `model_file` held an
  unrelated patch position, so replaying a demo from its states showed the can next to the patch.
  Images, states and actions are unchanged. `media/find_soda.gif` was re-rendered.

## Citation

```bibtex
@misc{wang2026rememberdidlearningbehavioral,
      title={Remember what you did?: Learning Behavioral Memories for Partially Observable Object Manipulation},
      author={Kuancheng Wang and Seungho Yeom and Jinglin Cao and Yuheng Zhi and Nikhil Shinde and Michael Yip},
      year={2026},
      eprint={2606.21188},
      archivePrefix={arXiv},
      primaryClass={cs.RO},
      url={https://arxiv.org/abs/2606.21188},
}
```

## License

MIT.