Datasetpublic

XXXXyu/libero-textual-memory-annotations

LIBERO predicate and textual-memory annotations This release contains predicate-derived annotations for the complete training split of LIBERO-Spatial, LIBERO-Object, LIBERO-Goal, and LIBERO-10. Its textual_memory field provides observation-grounded textual memory for each frame. Item Value Demonstrations 2,000 Frames 338,575 Demonstrations reaching simulator success 2,000 QA issues 0 Each JSONL row stores raw goal and auxiliary predicates plus independent… See the full description on the dataset page: https://huggingface.co/datasets/XXXXyu/libero-textual-memory-annotations.

sourceHugging Facecc-by-4.0updated 1mo agoView on Hugging Face

0likes28downloads

Dataset Card

LIBERO predicate and textual-memory annotations

This release contains predicate-derived annotations for the complete training split of LIBERO-Spatial, LIBERO-Object, LIBERO-Goal, and LIBERO-10. Its textual_memory field provides observation-grounded textual memory for each frame.

Item	Value
Demonstrations	2,000
Frames	338,575
Demonstrations reaching simulator success	2,000
QA issues	0

Each JSONL row stores raw goal and auxiliary predicates plus independent language fields for completed progress, remaining progress, the next demonstrated transition, and textual memory. textual_memory uses only stable goal predicates completed in the current observation; it does not use future frames.

The original LIBERO HDF5 demonstrations are not redistributed. Download them from the official LIBERO dataset to train the adapter or reconstruct images and robot state. The derived annotations use the official dataset's CC BY 4.0 license; retain LIBERO attribution when redistributing or modifying them.

The matching suffix-LoRA adapter is available at https://huggingface.co/xxxxyu/oxygen-pi05-textual-memory-lora.