Gated Memory Policy: In-context Memorization and Adaptation [CoRL 2026]
See the codeYihuai Gao, Jinyun Liu, Shuang Li, Shuran Song
Stanford University
Project Website, ArXiv, Models, Datasets
This repository contains source code for gated memory policy training, simulation data collection and evaluation (Memimic & RoboMimic and Mikasa-Robo benchmarks), and real-world robot deployment with in-the-wild checkpoints.
Major Update (Jun 29th, 2026): iPhUMI is released, huge shout out to Austin Patel! Please feel free to try the cup placement policy in-the-wild!
Tarined cups purchase list
Please refer to the policy codebase for instructions to set up the conda environment and serve the in-the-wild checkpoint.
For the real-world deployment codebase, please checkout the gated-memory-policy branch in case there are compatibility issues.
| Repo | What it does |
|---|---|
imitation-learning-policies/ | Policy training and inference serving |
real-env/ | Real-world robot deployment (Natively support UR5, ARX5) |
mujoco-env/ | MuJoCo sim, data collection and evaluation for the Memimic (Ours) and RoboMimic benchmarks |
mikasa-robo-env/ | ManiSkill sim, data collection and evaluation for the Mikasa-Robo benchmark |
Note: We use the submodule
real-envto manage the real-world robot deployment codebase (shared with other projects). Please usegit submodule update --init --recursiveto initialize the submodule.
We are grateful to the following amazing open-sourced projects that made this work possible:
If you find this work useful, please cite:
@misc{gao2026gatedmemorypolicy,
title = {Gated Memory Policy: In-Context Memorization and Adaptation},
author = {Yihuai Gao and Jeff Jinyun Liu and Shuang Li and Shuran Song},
year = {2026},
eprint = {2604.18933},
archivePrefix = {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2604.18933},
}
Gated Memory Policy: In-context Memorization and Adaptation [CoRL 2026]
See the codeYihuai Gao, Jinyun Liu, Shuang Li, Shuran Song
Stanford University
Project Website, ArXiv, Models, Datasets
This repository contains source code for gated memory policy training, simulation data collection and evaluation (Memimic & RoboMimic and Mikasa-Robo benchmarks), and real-world robot deployment with in-the-wild checkpoints.
Major Update (Jun 29th, 2026): iPhUMI is released, huge shout out to Austin Patel! Please feel free to try the cup placement policy in-the-wild!
Tarined cups purchase list
Please refer to the policy codebase for instructions to set up the conda environment and serve the in-the-wild checkpoint.
For the real-world deployment codebase, please checkout the gated-memory-policy branch in case there are compatibility issues.
| Repo | What it does |
|---|---|
imitation-learning-policies/ | Policy training and inference serving |
real-env/ | Real-world robot deployment (Natively support UR5, ARX5) |
mujoco-env/ | MuJoCo sim, data collection and evaluation for the Memimic (Ours) and RoboMimic benchmarks |
mikasa-robo-env/ | ManiSkill sim, data collection and evaluation for the Mikasa-Robo benchmark |
Note: We use the submodule
real-envto manage the real-world robot deployment codebase (shared with other projects). Please usegit submodule update --init --recursiveto initialize the submodule.
We are grateful to the following amazing open-sourced projects that made this work possible:
If you find this work useful, please cite:
@misc{gao2026gatedmemorypolicy,
title = {Gated Memory Policy: In-Context Memorization and Adaptation},
author = {Yihuai Gao and Jeff Jinyun Liu and Shuang Li and Shuran Song},
year = {2026},
eprint = {2604.18933},
archivePrefix = {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2604.18933},
}