Add files using upload-large-folder tool

README.md

@@ -1,3 +1,134 @@
----
-license: apache-2.0
----
+# World Pilot: Steering Vision-Language-Action Models with World-Action Priors
+
+This repository hosts the released **WorldPilot** checkpoint for the **LIBERO** setting from the open-source WorldPilot project.
+
+Resources:
+
+- Project website: https://world-pilot.github.io/
+- Project code: https://github.com/ZefuLin/WorldPilot
+- Precomputed LIBERO cache: https://huggingface.co/datasets/Chedan86/WorldPilot-LIBERO-precompute
+
+## Overview
+
+WorldPilot steers a vision-language-action policy with priors from a world-action model. This release provides the public LIBERO checkpoint together with the run artifacts needed to load it from the public WorldPilot codebase.
+
+## Included Files
+
+```text
+checkpoints/steps_50000_pytorch_model.pt
+config.yaml
+dataset_statistics.json
+run_libero_train.sh
+summary.jsonl
+training.log
+training_log.jsonl
+source_code/Augur/...
+```
+
+- `checkpoints/steps_50000_pytorch_model.pt`: released WorldPilot checkpoint.
+- `config.yaml`: saved training configuration for this run.
+- `dataset_statistics.json`: statistic file saved with the WorldPilot run.
+- `run_libero_train.sh`: training launch script snapshot saved with the run.
+- `summary.jsonl`, `training.log`, `training_log.jsonl`: training summaries and logs.
+- `source_code/Augur/...`: source snapshot captured with the run for reference and provenance.
+
+## Intended Use
+
+This release is intended for:
+
+1. loading the released WorldPilot checkpoint with the public WorldPilot codebase;
+2. running public LIBERO-plus evaluation;
+3. inspecting the released configuration and run artifacts.
+
+This repository is not a standalone full dependency bundle. Public training and evaluation still require the upstream assets and environments documented in the main project repository.
+
+## Using This Checkpoint
+
+The public evaluation entrypoints are:
+
+- `examples/LIBERO-plus/eval_files/eval_libero_single.sh`
+- `examples/LIBERO-plus/eval_files/eval_libero_batch.sh`
+
+Set:
+
+```bash
+WORLDPILOT_CKPT=/path/to/WorldPilot-LIBERO/checkpoints/steps_50000_pytorch_model.pt
+```
+
+The public evaluation scripts also require:
+
+- a Cosmos Policy checkpoint;
+- a Cosmos dataset statistics JSON;
+- a Cosmos T5 embeddings pickle;
+- a valid `~/.libero-plus/config.yaml`.
+
+If you want the evaluation script to use the bundled source snapshot, set:
+
+```bash
+USE_CHECKPOINT_SOURCE=true
+```
+
+The script will then add `source_code/` from this release to `PYTHONPATH`.
+
+Important:
+
+- `dataset_statistics.json` in this repository is the statistic file saved with the WorldPilot run.
+- It does not replace the `COSMOS_STATS` file required by the Cosmos server in the evaluation scripts.
+
+For the full setup, follow the public project documentation:
+
+- https://world-pilot.github.io/
+- https://github.com/ZefuLin/WorldPilot/tree/main/doc
+
+## Training Reference
+
+The public training path in the code release uses:
+
+- `examples/LIBERO/train_files/WorldPilot.yaml`
+- `examples/LIBERO/train_files/run_libero_train.sh`
+
+Training also depends on the upstream assets listed in the public docs, including:
+
+- `nvidia/Cosmos-Policy-LIBERO-Predict2-2B`
+- `facebook/VGGT-1B`
+- `StarVLA/Qwen3-VL-4B-Instruct-Action`
+- `amap_cvlab/ABot-M0-Pretrain`
+
+Training reads precomputed Cosmos cache from:
+
+```text
+datasets.vla_data.cosmos_cache_dir
+```
+
+You can reuse the published cache here:
+
+- https://huggingface.co/datasets/Chedan86/WorldPilot-LIBERO-precompute
+
+The included `config.yaml` and `run_libero_train.sh` are run artifacts for inspection and reproduction support. For active training, use the checked-out public repository and its current training docs.
+
+## Provenance and Terms
+
+The main WorldPilot code repository is released under Apache-2.0:
+
+- https://github.com/ZefuLin/WorldPilot
+
+This release also depends on upstream projects and assets that keep their own licenses and usage terms. Upstream checkpoints, datasets, and third-party components are not relicensed by this model card.
+
+The bundled `source_code/` snapshot is included for reference and provenance only. It does not supersede original upstream notices or license terms.
+
+## Citation
+
+If WorldPilot helps your research, please cite:
+
+```bibtex
+@article{worldpilot2026,
+  title={World Pilot: Steering Vision-Language-Action Models with World-Action Priors},
+  author={Anonymous},
+  journal={arXiv preprint arXiv:xxxx.xxxxx},
+  year={2026}
+}
+```
+
+## Acknowledgements
+
+We thank the teams behind ABot-Manipulation, cosmos-policy, LIBERO, LIBERO-plus, and LeRobot.

checkpoints/steps_50000_pytorch_model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7fbed14a2c239fc830748d568f4d7452c288c1c55c7dee84337ff0986ba5ba9
+size 12692203808

dataset_statistics.json

@@ -0,0 +1,133 @@
+{
+  "franka": {
+    "action": {
+      "mean": [
+        0.06413520714984491,
+        0.08330044630341807,
+        -0.09768853995662471,
+        3.989756358071042e-05,
+        0.0074605874788875775,
+        -0.0023408993505514594,
+        0.5327441073380985
+      ],
+      "std": [
+        0.34435420775153375,
+        0.37664413619011156,
+        0.4530659887664812,
+        0.040994048642083246,
+        0.06561666470959421,
+        0.08020121465404138,
+        0.4990030574544087
+      ],
+      "max": [
+        0.9375,
+        0.9375,
+        0.9375,
+        0.3557142913341522,
+        0.375,
+        0.375,
+        1.0
+      ],
+      "min": [
+        -0.9375,
+        -0.9375,
+        -0.9375,
+        -0.2582142949104309,
+        -0.375,
+        -0.3675000071525574,
+        0.0
+      ],
+      "q01": [
+        -0.8785714507102966,
+        -0.8758928775787354,
+        -0.9375,
+        -0.1510714292526245,
+        -0.20678570866584778,
+        -0.2742857038974762,
+        0.0
+      ],
+      "q99": [
+        0.9375,
+        0.9107142686843872,
+        0.9375,
+        0.20357142388820648,
+        0.26357144117355347,
+        0.375,
+        1.0
+      ],
+      "mask": [
+        true,
+        true,
+        true,
+        true,
+        true,
+        true,
+        false
+      ]
+    },
+    "state": {
+      "mean": [
+        -0.051236557416044756,
+        0.03320404933765531,
+        0.7923003721695681,
+        2.9629008036393385,
+        -0.15774205785531267,
+        -0.13519181454410922,
+        0.02689367747650697,
+        -0.027077807328448854
+      ],
+      "std": [
+        0.10707255982690211,
+        0.14929271152454146,
+        0.37628407735102676,
+        0.3693053742611154,
+        0.8810664091949664,
+        0.33025891946901687,
+        0.01434538622885222,
+        0.01422763564543161
+      ],
+      "max": [
+        0.21031762659549713,
+        0.39128610491752625,
+        1.3660105466842651,
+        3.6714255809783936,
+        3.560650587081909,
+        1.386339545249939,
+        0.04233968257904053,
+        0.0013633022317662835
+      ],
+      "min": [
+        -0.4828203022480011,
+        -0.3255046010017395,
+        0.008128180168569088,
+        0.35277295112609863,
+        -3.641430377960205,
+        -1.842738389968872,
+        -0.0013586411951109767,
+        -0.042040832340717316
+      ],
+      "q01": [
+        -0.42401049643754957,
+        -0.2838300323486328,
+        0.009925739830359817,
+        1.3085840785503386,
+        -2.886677579879761,
+        -1.1599004411697387,
+        0.001503719249740243,
+        -0.040336399003863335
+      ],
+      "q99": [
+        0.1530261474847791,
+        0.3629165390133857,
+        1.2910678112506866,
+        3.303542451858519,
+        2.7496529006957933,
+        0.6893712210655194,
+        0.040610933862626555,
+        -0.0015016929572448147
+      ]
+    },
+    "num_transitions": 272104,
+    "num_trajectories": 1693
+  }
+}