# robomimic: Demonstration train-validation split

Add training and validation trajectory filter keys to a selected HDF5 demonstration dataset. Inspect the pinned source and setup before use.

ID: `robomimic-robomimic-scripts-split-train-val-py`

Record version: `source-8e5cc64eed375b20991e5261901638415f44868a4fad124ebb318d39c7c16a5a`

Hub commit: `703369d7a095afb46da723634a9d983d2eda3054`

## Access

Fetch the pinned repository, verify this source checksum, and use upstream installation and invocation. This record does not install or execute code.

Optional inert file retrieval: `python /PATH_TO_SKILL/scripts/rsh.py fetch-source robomimic-robomimic-scripts-split-train-val-py --project ./experiment --expected-source-revision d309eaecc18acf4152a830a895a6984b8ac71b05 --expected-source-sha256 8e5cc64eed375b20991e5261901638415f44868a4fad124ebb318d39c7c16a5a --json`. This verifies one file, preserves provenance and never installs or runs it. See [source-file retrieval](https://hub.superfastrobotics.com/source-files.md).

Hub preparation adapter: **false**. Runtime verification: **not_run**.

Upstream setup: https://github.com/ARISE-Initiative/robomimic/blob/d309eaecc18acf4152a830a895a6984b8ac71b05/README.md

## Source and license

- Repository: https://github.com/ARISE-Initiative/robomimic
- Pinned source: https://github.com/ARISE-Initiative/robomimic/blob/d309eaecc18acf4152a830a895a6984b8ac71b05/robomimic/scripts/split_train_val.py
- Revision: `d309eaecc18acf4152a830a895a6984b8ac71b05`
- Path: `robomimic/scripts/split_train_val.py`
- Source SHA256: `8e5cc64eed375b20991e5261901638415f44868a4fad124ebb318d39c7c16a5a`
- Source bytes: 3566
- Original authors: Ajay Mandlekar, Danfei Xu, Josiah Wong, Soroush Nasiriany, Chen Wang, Matthew Bronars, Vaibhav Saxena
- License: [MIT](https://github.com/ARISE-Initiative/robomimic/blob/d309eaecc18acf4152a830a895a6984b8ac71b05/LICENSE)
- License scope: Selected first-party workflow Python source only; preserve all upstream copyright, author and per-file notices. This does not relicense model weights, datasets, Gemma or other third-party model code, robot firmware, simulator binaries, meshes, dependencies or external services; their separate terms and access requirements remain applicable.

License SHA256: `7cdbfab482b23a4d925d59ff169ab0bc5f8c97ceb0db79f9fd5bf46ef8aa1556`

## Artifact

```json
{
  "kind": "workflow",
  "format": "python",
  "workflow_role": "dataset-preparation"
}
```

## Requirements

```json
{
  "python": "Python >=3 as declared by pinned setup.py; follow the pinned installation guide for its supported environment",
  "packages": [
    "Pinned robomimic checkout",
    "PyTorch, h5py and dependencies declared by pinned setup.py",
    "Matching robosuite/MuJoCo environment when the selected workflow uses simulation"
  ],
  "display": "Training/evaluation requires an explicitly selected policy configuration, compatible HDF5 demonstration dataset and, for rollout evaluation, the matching simulator and checkpoint. Dataset tools may modify HDF5 files in place.",
  "execution_domain": "computation",
  "source_imports": [
    "argparse",
    "h5py",
    "numpy",
    "robomimic.utils.file_utils"
  ]
}
```

## Invocation

```json
{
  "entrypoint": "python robomimic/scripts/split_train_val.py",
  "notes": "Follow pinned robomimic setup and provide the required --config/--algo, dataset, checkpoint and environment options for this script. Conversion and split tools may modify the input dataset; review their source first. The example training loops are instructional; no policy performance or robot compatibility has been measured."
}
```

## Compatibility

```json
[]
```

Missing environment fields remain unknown. Source discovery labels do not establish robot compatibility.

## Evidence

Level: **source_reviewed**. Runtime verification: **not_run**.

Verified exact source, applicable license and author/setup document bytes at a full Git commit; parsed the Python AST and reviewed this guarded operational entrypoint and its explicit license notices. No upstream code, installation, robot command, dataset, mesh or checkpoint was executed or loaded.

Source review and contributor reports do not establish task completion. Preserve upstream attribution and licenses.