---
language:
- en
tags:
- robotics
- vision-language-action
- action-chunking
- chunktrust
library_name: openpi
---
# ChunkTrust

**Adapting Execution Horizons for Robot Policies with Action-Expert Evidence**

[Paper](https://arxiv.org/abs/2609.39754) · [Project Page](https://hf618.github.io/ChunkTrust.github.io/) · [Code](https://github.com/hf618/ChunkTrust)

Public model assets for ChunkTrust. The maintained implementation, installation instructions, and training and evaluation recipes are in the [code repository](https://github.com/hf618/ChunkTrust).

## Data and checkpoints

Training data for the simulation experiments come from [RoboTwin 2.0](https://github.com/RoboTwin-Platform/RoboTwin) and [RoboCasa GR1 Tabletop](https://github.com/robocasa/robocasa-gr1-tabletop-tasks). The QHA protocols use RoboTwin 2.0 clean50 demonstrations. Demonstration datasets and preprocessed teacher caches are not included in this repository.

| Checkpoint group | Step | Published assets |
|---|---:|---|
| pi0 RoboTwin 2.0 task policies | 30000 | 8 task-specific policies with normalization statistics |
| pi0.5 RoboTwin 2.0 task policies | 20000 | 8 task-specific policies with normalization statistics |
| pi0.5 RoboTwin 2.0 multitask policy | 30000 | 50-task policy with normalization statistics |
| pi0 QHA, eight-task augmentation | 5000 | Adapter and configuration |
| pi0.5 QHA, eight-task augmentation | 5000 | Adapter and configuration |
| pi0.5 QHA, six-task training and two held-out tasks | 10000 | Adapter and configuration |

QHA adapters must be used with the matching frozen base policy and normalization settings. See the [training recipes](https://github.com/hf618/ChunkTrust/blob/main/docs/training.md) and [model guide](https://github.com/hf618/ChunkTrust/blob/main/docs/models.md) for the checkpoint-to-protocol mapping.

The pi0.5 RoboCasa GR1 policy is hosted separately on [ModelScope](https://modelscope.cn/models/NewGain/pi05_robocasa_gr1_tabletop24).

## Download

The [asset catalog](code/assets.json) lists all published checkpoint files at pinned revisions. From a clone of the code repository:

```bash
curl -L https://huggingface.co/Niugan/ChunkTrust/resolve/main/code/assets.json -o assets.json
python scripts/download_asset.py qha_heldout --catalog assets.json --destination checkpoints/qha-heldout/10000
```

Other adapter keys are `qha_pi0_8task_step5000` and `qha_pi05_8task_step5000`. For base policies, use the task or multitask model key in the catalog.

## Repository contents

- `checkpoints/pi0/`: pi0 task-specific policies.
- `checkpoints/pi05/`: pi0.5 task-specific and multitask policies.
- `checkpoints/qha_*/`: the three QHA adapters.
- `code/`: public source snapshot and download catalog.
- `media/`: introduction video and cover image.

[Download the public source snapshot](code/ChunkTrust-source-0.1.0.tar.gz), corresponding to GitHub commit `27cfab89a82295cc534e9dd641bdec7c85499797`. The [4K introduction is also available on GitHub](https://github.com/hf618/ChunkTrust/releases/tag/media-v1).

Base-model and dataset licenses apply independently of the code repository's MIT license.