yinchenghust/robomemarena_lerobot

SimpleMemVLA Dataset

This repository is part of the SimpleMemVLA project, presented in the paper SimpleMemVLA: A Simple but Effective Native-Video Memory for Vision-Language-Action Models.

• Paper: arXiv:2609.05533
• Code: GitHub - wadeKeith/SimpleMemVLA
• Project page: Hugging Face collection

This repository hosts datasets in LeRobot v3 format for training and evaluating vision-language-action models with long-term memory capabilities. The datasets cover benchmark suites including RMBench, RoboMME, MIKASA-Robo, RoboMemArena, and LIBERO, and are available on Hugging Face under yinchenghust/<benchmark>_lerobot.

Sample Usage

huggingface-cli download yinchenghust/libero_lerobot --repo-type dataset --local-dir data/datasets/yinchenghust/libero_lerobot

For full training, evaluation, and benchmark details, see the GitHub repository. This work is released under the MIT License.