mikasa-robo/mikasa-robo-vla-lerobot
MIKASA-Robo-VLA MIKASA-Robo-VLA is a memory-intensive robotic manipulation benchmark for Vision-Language-Action (VLA) research. It contains 90 language-conditioned manipulation tasks and 22,500 oracle trajectories across 10 memory types, available in RLDS and LeRobotDataset v3 formats. Paper: MIKASA-Robo-VLA: Benchmarking Memory in VLA Models for Long-Horizon Manipulation Project page: https://mikasarobo.github.io/ Code: https://github.com/CognitiveAISystems/MIKASA-Robo… See the full description on the dataset page: https://huggingface.co/datasets/mikasa-robo/mikasa-robo-vla-lerobot.
| Author / organization | mikasa-robo |
|---|---|
| License (as declared) | mit |
| Downloads | 898 (as reported by Hugging Face) |
| Likes | 2 |
| Created | 20 April 2026 |
| Last updated at source | 16 October 2026 |
Popularity, last 2 days
898 now (+6 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
Recent changes
- 16 Oct 2026 Description: none → MIKASA-Robo-VLA MIKASA-Robo-VLA is a memory-intensive robotic manipulation benchmark for Vision-Language-Action (VLA) research. It contains 90 language-conditioned manipulation tasks and 22,500 oracle trajectories across 10 memory types, av
- 16 Oct 2026 Topics added task_categories:robotics, memory, manipulation, vision-language-action, benchmark
Changes detected by Silytics between refreshes of the source.
Related papers
More from mikasa-robo
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.