Instructions to use WitneyWW/acts-sensoronly-ablation with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Cosmos
How to use WitneyWW/acts-sensoronly-ablation with Cosmos:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Download preprocessing/make_manifest.py from WitneyWW/acts-sensoronly-ablation: direct link, hf CLI and curl.
- Browser
- Download file 1.98 kB
-
https://huggingface.co/WitneyWW/acts-sensoronly-ablation/resolve/refs%2Fpr%2F2/preprocessing/make_manifest.py
- Command line
-
hf download hf://WitneyWW/acts-sensoronly-ablation@refs/pr/2/preprocessing/make_manifest.py
-
curl -L -o make_manifest.py https://huggingface.co/WitneyWW/acts-sensoronly-ablation/resolve/refs%2Fpr%2F2/preprocessing/make_manifest.py
1.98 kB
| """Write the encoder manifest (episodes.jsonl) for one task from the downloaded raw/ tree. | |
| python preprocessing/make_manifest.py --raw-root $ACTS_DATA_ROOT/raw --task rope --out OUT/episodes_rope.jsonl | |
| raw/ layout (WitneyWW/vtm-cosmos-data): | |
| raw/videos/<episode>/{view_middle,tactile_left,tactile_right,wrist_left,wrist_right}.mp4 (30 fps) | |
| raw/parquet/<episode>.parquet per-native-frame sensor poses + calibrated normal force | |
| raw/episodes.json [{episode, tree, num_frames}] for the 166 split episodes | |
| Row schema = what third_party/wan22_vae/encode_tactile_wan22.py reads.""" | |
| from __future__ import annotations | |
| import argparse | |
| import json | |
| from pathlib import Path | |
| TREES = {"umi_all": "mixed_mb", "rope": "mixed_rope", "pusht": "mixed_pushT"} | |
| STREAMS = ["view_middle", "tactile_left", "tactile_right", "wrist_left", "wrist_right"] | |
| def main(): | |
| ap = argparse.ArgumentParser() | |
| ap.add_argument("--raw-root", type=Path, required=True) | |
| ap.add_argument("--task", required=True, choices=sorted(TREES)) | |
| ap.add_argument("--out", type=Path, required=True) | |
| ap.add_argument("--episodes", nargs="*", default=None, help="subset (episode stems)") | |
| a = ap.parse_args() | |
| rows = [r for r in json.loads((a.raw_root / "episodes.json").read_text()) if r["tree"] == TREES[a.task]] | |
| if a.episodes: | |
| rows = [r for r in rows if r["episode"] in set(a.episodes)] | |
| out = [] | |
| for r in rows: | |
| d = a.raw_root / "videos" / r["episode"] | |
| row = {"episode": r["episode"], "num_frames": int(r["num_frames"]), "task": a.task} | |
| for s in STREAMS: | |
| p = d / f"{s}.mp4" | |
| if not p.is_file(): | |
| raise FileNotFoundError(p) | |
| row[s] = str(p.resolve()) | |
| out.append(row) | |
| a.out.parent.mkdir(parents=True, exist_ok=True) | |
| a.out.write_text("".join(json.dumps(r) + "\n" for r in out)) | |
| print(f"{len(out)} episodes -> {a.out}") | |
| if __name__ == "__main__": | |
| main() | |