DiffOG

KO | EN

gitlite — search

#javascript #react #ai #llm #go #deep-learning #docker #macos #java #awesome #kubernetes #security #mcp #openai

(★ 21)

(T-RO 2025) DiffOG: Differentiable Policy Trajectory Optimization with Generalizability

DiffOG 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# DiffOG: Differentiable Policy Trajectory Optimization with Generalizability

_Published in IEEE Transactions on Robotics_

[Project](https://zhengtongxu.github.io/diffog-website/) | [arXiv](https://arxiv.org/abs/2504.13807) | [Video](https://drive.google.com/file/d/19vxJWG5CZg6uCrhsZEANto7W33nWLZbl/view)

DiffOG is a transformer-based differentiable trajectory optimization framework that enhances visuomotor policies by refining action trajectories to be smoother and constraint-compliant while maintaining fidelity to the demonstration distribution.

![DiffOG](https://raw.githubusercontent.com/ZhengtongXu/DiffOG/main/teaser/first_page.png)

## Installation

```console
$ cd diffog
$ mamba env create -f conda_environment.yaml
$ bash install_third_party.sh
```

The environment also requires MuJoCo. Please refer to the [MuJoCo documentation](https://mujoco.readthedocs.io/en/2.1.3/overview.html) for installation. Make sure to set the following environment variables before running:
```console
$ export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:${HOME}/.mujoco/mujoco210/bin
$ export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/lib/nvidia
$ export MUJOCO_GL=egl
```

## Dataset

To run DiffOG on [robomimic](https://robomimic.github.io/) tasks, download and extract the data under the `data/` directory:

```console
$ cd diffog
$ wget https://diffusion-policy.cs.columbia.edu/data/training/robomimic_image.zip
$ unzip robomimic_image.zip -d data/
$ rm robomimic_image.zip
```

## Training

Activate the conda environment:
```console
$ conda activate diffogenv
```

All training logs are uploaded to [wandb](https://wandb.ai). Login to wandb (if you haven't already):
```console
$ wandb login
```

### DiffOG Dataset Training

In this mode, supervised learning is performed directly on the demonstration dataset.

For example, train DiffOG on the **Tool Hang** task:
```console
$ python train.py --config-dir=./diffog/config --config-name=train_diffusion_diffog_tool.yaml hydra.run.dir=data/outputs/train_diffusion_diffog_tool
```

### DiffOG Refine Training

In this mode, supervised learning refines the actions generated by a pre-trained base policy. This requires a pretrained base diffusion policy checkpoint. Specify the path to the checkpoint via the `init_path` field in the YAML config before running.

For example, train DiffOG Refine on the **Tool Hang** task:
```console
$ python train.py --config-dir=./diffog/config --config-name=train_diffusion_diffog_tool_res_ddim.yaml hydra.run.dir=data/outputs/train_diffusion_diffog_tool_res_ddim
```

Note that each job needs its own output directory. You don't need to create the folder manually; the system will automatically create a new folder for you when you run the job with a different `hydra.run.dir`.

## Differentiable Optimization Layer

The core differentiable optimization layer is implemented in `diffog/model/diffog_layer.py`. It contains two main components:

- **`DiffogTf`**: The differentiable optimization module that formulates and solves a Quadratic Program to enforce smoothness and feasibility constraints on the predicted action trajectory. It uses a Transformer encoder (`TemporalTransformerEncoder`) to parameterize the learnable cost matrix **Q**, which is constructed as a symmetric positive definite matrix via Cholesky factorization. The QP is solved via [qpth](https://github.com/locuslab/qpth) for differentiable optimization. Key parameters include:
  - `constraint`: the bound on consecutive action differences
  - `smooth_weight`: weight for the smoothness regularization term

- **`TemporalTransformerEncoder`**: A Transformer-based encoder that processes the temporal action sequence and outputs a single embedding vector used to construct the input-dependent **Q** matrix.

The `forward` method enforces inter-step smoothness constraints during training, while `forward_with_past_action` additionally constrains the first predicted action to be close to the most recent executed action, ensuring temporal consistency during inference.

## BibTeX

If you find this codebase useful, consider citing:

```bibtex
@article{xu2025diffog,
  title={DiffOG: Differentiable Policy Trajectory Optimization with Generalizability},
  author={Xu, Zhengtong and Miao, Zichen and Qiu, Qiang and Zhang, Zhe and She, Yu},
  journal={IEEE Transactions on Robotics},
  year={2025},
  publisher={IEEE}
}
```

## Acknowledgement

This repository, during construction, referenced the code structure of [diffusion policy](https://github.com/real-stanford/diffusion_policy). We sincerely thank the authors of diffusion policy for open-sourcing such an elegant codebase!

DiffOG

DiffOG 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

DiffOG

DiffOG 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README