dataset_id large_stringclasses 5
values | path large_stringclasses 5
values | domain large_stringclasses 1
value | embodiment_id large_stringclasses 1
value | instruction_slug large_stringclasses 5
values | instruction large_stringclasses 5
values | lerobot_version large_stringclasses 1
value | robot_type large_stringclasses 1
value | license large_stringclasses 1
value | capture_fingerprint large_stringclasses 1
value | source_uri large_stringclasses 1
value | source_revision large_stringclasses 1
value | source_license large_stringclasses 1
value | status large_stringclasses 1
value | total_episodes int64 100 100 | total_frames int64 44k 45.3k | duration_s float64 881 906 | fps float64 50 50 | storage_bytes int64 55.3M 58.2M | camera_keys large listlengths 3 3 | split_counts large_stringclasses 1
value | quality_counts large_stringclasses 1
value | outcome_label_coverage float64 0 0 | added_at large_stringclasses 1
value | updated_at large_stringclasses 1
value | related_datasets large listlengths 0 0 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
qualia-dataset-sim/openarm/place-the-blue-cube-in-the-box | data/openarm/place-the-blue-cube-in-the-box | sim | openarm | place-the-blue-cube-in-the-box | place the blue cube in the box | v2.1 | openarm_bimanual_sim | cc-by-nc-sa-4.0 | sha256:255b2c185de9ba75677ea3f170d1da79676fb9701294d99e0349b45e7df6c64d | https://huggingface.co/datasets/qualia-robotics/openarm-cube-distractor-lang-v3 | e71060d87b08a76cda9efb58126e1dc61a0046dc | apache-2.0 | unreviewed | 100 | 44,367 | 887.34 | 50 | 56,261,134 | [
"observation.images.base",
"observation.images.wrist_left",
"observation.images.wrist_right"
] | {"train":100} | {"unreviewed":100} | 0 | unknown | unknown | [] |
qualia-dataset-sim/openarm/place-the-green-cube-in-the-box | data/openarm/place-the-green-cube-in-the-box | sim | openarm | place-the-green-cube-in-the-box | place the green cube in the box | v2.1 | openarm_bimanual_sim | cc-by-nc-sa-4.0 | sha256:255b2c185de9ba75677ea3f170d1da79676fb9701294d99e0349b45e7df6c64d | https://huggingface.co/datasets/qualia-robotics/openarm-cube-distractor-lang-v3 | e71060d87b08a76cda9efb58126e1dc61a0046dc | apache-2.0 | unreviewed | 100 | 44,038 | 880.76 | 50 | 55,309,322 | [
"observation.images.base",
"observation.images.wrist_left",
"observation.images.wrist_right"
] | {"train":100} | {"unreviewed":100} | 0 | unknown | unknown | [] |
qualia-dataset-sim/openarm/place-the-purple-cube-in-the-box | data/openarm/place-the-purple-cube-in-the-box | sim | openarm | place-the-purple-cube-in-the-box | place the purple cube in the box | v2.1 | openarm_bimanual_sim | cc-by-nc-sa-4.0 | sha256:255b2c185de9ba75677ea3f170d1da79676fb9701294d99e0349b45e7df6c64d | https://huggingface.co/datasets/qualia-robotics/openarm-cube-distractor-lang-v3 | e71060d87b08a76cda9efb58126e1dc61a0046dc | apache-2.0 | unreviewed | 100 | 45,276 | 905.52 | 50 | 57,600,432 | [
"observation.images.base",
"observation.images.wrist_left",
"observation.images.wrist_right"
] | {"train":100} | {"unreviewed":100} | 0 | unknown | unknown | [] |
qualia-dataset-sim/openarm/place-the-red-cube-in-the-box | data/openarm/place-the-red-cube-in-the-box | sim | openarm | place-the-red-cube-in-the-box | place the red cube in the box | v2.1 | openarm_bimanual_sim | cc-by-nc-sa-4.0 | sha256:255b2c185de9ba75677ea3f170d1da79676fb9701294d99e0349b45e7df6c64d | https://huggingface.co/datasets/qualia-robotics/openarm-cube-distractor-lang-v3 | e71060d87b08a76cda9efb58126e1dc61a0046dc | apache-2.0 | unreviewed | 100 | 44,491 | 889.82 | 50 | 56,801,200 | [
"observation.images.base",
"observation.images.wrist_left",
"observation.images.wrist_right"
] | {"train":100} | {"unreviewed":100} | 0 | unknown | unknown | [] |
qualia-dataset-sim/openarm/place-the-yellow-cube-in-the-box | data/openarm/place-the-yellow-cube-in-the-box | sim | openarm | place-the-yellow-cube-in-the-box | place the yellow cube in the box | v2.1 | openarm_bimanual_sim | cc-by-nc-sa-4.0 | sha256:255b2c185de9ba75677ea3f170d1da79676fb9701294d99e0349b45e7df6c64d | https://huggingface.co/datasets/qualia-robotics/openarm-cube-distractor-lang-v3 | e71060d87b08a76cda9efb58126e1dc61a0046dc | apache-2.0 | unreviewed | 100 | 44,279 | 885.58 | 50 | 58,231,765 | [
"observation.images.base",
"observation.images.wrist_left",
"observation.images.wrist_right"
] | {"train":100} | {"unreviewed":100} | 0 | unknown | unknown | [] |
Qualia Simulation Datasets
| Release | release-2026-07-19.3 (changelog) |
| Domain | Robot simulation |
| Format | LeRobot v2.1 |
| Access | Public |
| Contact | Start a discussion |
This collection is public. Provenance and source licensing are recorded in each dataset's
dataset.yaml.
Dataset description
This repository collects Qualia's simulated robot datasets in one place. Each dataset contains one language instruction and records the source revision used to prepare it.
The collection will grow as new simulation runs become available. A change to the instruction, robot, cameras, simulator setup or control space creates a new dataset; compatible episodes may be appended after validation.
Using the collection
The Hub viewer opens catalog.parquet, which is the quickest way to see what is available. The
trajectory files are stored in the dataset directories linked below.
This repository is not a single LeRobot or OpenPI training root. Choose one dataset directory, or prepare a separate training dataset from the recordings you need. Pin the release or commit used for each experiment.
Versioning
main contains the newest accepted state of the collection. Dated release tags identify reviewed
snapshots; training manifests should record both the release tag and full commit SHA. The collection
release is separate from LeRobot v2.1, which describes the on-disk data format.
Included datasets
All five datasets use the simulated bimanual OpenArm setup and come from
qualia-robotics/openarm-cube-distractor-lang-v3.
| Dataset | Instruction | Episodes | Frames | Recording time |
|---|---|---|---|---|
| Red cube | place the red cube in the box | 100 | 44,491 | 14m 50s |
| Green cube | place the green cube in the box | 100 | 44,038 | 14m 41s |
| Blue cube | place the blue cube in the box | 100 | 44,367 | 14m 47s |
| Yellow cube | place the yellow cube in the box | 100 | 44,279 | 14m 46s |
| Purple cube | place the purple cube in the box | 100 | 45,276 | 15m 6s |
Dataset structure
data/
openarm/
<instruction>/
dataset.md # datasheet for people
dataset.yaml # source, simulator and capture metadata
meta/ # LeRobot v2.1 metadata
data/ # one Parquet file per episode
videos/ # one video per episode and camera
qualia/ # episode provenance, statistics and validation
Each directory under data/openarm/<instruction>/ is a complete, one-instruction LeRobot v2.1
dataset. The root catalog configuration exposes only catalog.parquet; it deliberately does not
combine the trajectory tables.
Dataset creation
Curation rationale
Keeping instructions separate makes it possible to inspect, review and sample them independently. It also keeps source lineage and simulator metadata close to the episodes instead of relying on informal links.
Source data and processing
This release converts five instructions from one LeRobot v3.0 source dataset. The source repository was not modified. Its episodes were grouped by task, converted to v2.1, and given new local episode and frame indices while preserving their source identities. All statistics were recalculated from the converted files.
Validation
Every dataset is checked for task consistency, episode and frame indices, timestamps, feature shapes, camera metadata, missing files, non-finite values and stale statistics. Video frame counts, durations and frame rates were also checked for this release.
Collection statistics
| Datasets | Embodiments | Episodes | Frames | Recording time | Storage |
|---|---|---|---|---|---|
| 5 | 1 | 500 | 222,451 | 1h 14m | 271.0 MiB |
By embodiment
| Embodiment | Datasets | Episodes | Frames | Recording time | Storage |
|---|---|---|---|---|---|
| OpenArm | 5 | 500 | 222,451 | 1h 14m | 271.0 MiB |
Action and state ranges are calculated separately for each dataset because robots and control setups may use different action spaces.
More charts
Known limitations
- Episode-level success, termination and simulator seeds were not available in the source dataset.
- The episodes have not yet been reviewed individually;
unreviewedis not a success or failure label. - Simulator version, asset revisions, camera mounting and action units are currently unknown.
- All five instructions come from one cube-placement setup, so the collection does not yet represent broad task or scene diversity.
Contributing
We welcome new simulation tasks, additional episodes, corrected metadata and episode-level review labels. See CONTRIBUTING.md for the required files and checks.
License and provenance
Qualia Robotics releases this collection under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license. Reuse requires attribution, must be non-commercial, and adaptations must use the same license.
The original simulation source is recorded as Apache-2.0 in the manifests. Provenance is retained
per episode, and the exact source revision is recorded in every dataset.yaml.
- Downloads last month
- 92