README.md

---
license: "mit"
datasets:
- robomimic
- Push-T
arxiv: 2307.01849
tags:
- robotics
- diffusion
- self-supervised learning
- behaviour-cloning
- diffusion policy
---
# Crossway Diffusion: Improving Diffusion-based Visuomotor Policy via Self-supervised Learning

This is the checkpoint release of the [ICRA'24](https://2024.ieee-icra.org/) paper [Crossway Diffusion: Improving Diffusion-based Visuomotor Policy via Self-supervised Learning](https://arxiv.org/abs/2307.01849). 


This repo contains:
- Pretrained models (500 epochs): `./checkpoints/crossway_diffusion-<task>-s<seed>.ckpt`
- SHA1 checksums: `./checkpoints/crossway_diffusion-<task>-s<seed>.ckpt.sha1`
- Evaluation results (1000 episodes): `./checkpoints/crossway_diffusion-<task>-s<seed>.ckpt.json`

Available tasks: `can_mh`, `can_ph`, `lift_mh`, `lift_ph`, `square_mh`, `square_ph`, `transport_mh`, `transport_ph`, `tool_hang_ph`, `pusht`

Available seeds: `42`, `43`, `44`

Please replace `<task>` with a specific task name and replace `<seed>` with `42`, `43`, or `44` to navigate the repo.
For example, the path of pretrained model for task Push-T trained on seed 42 is `./checkpoints/crossway_diffusion-pusht-s42.ckpt`

For the training and evaluation code, please visit our [github repo](https://github.com/LostXine/crossway_diffusion).

Watch [🎬 video presentation](https://youtu.be/9deKHueZBuk)!

For questions, please raise a [github issue](https://github.com/LostXine/crossway_diffusion/issues).