This dataset was used to train a VLM model on robotic planning. It should be used in conjunction with the Places365 dataset (http://places.csail.mit.edu)
Dataset Details
Dataset Description
For each sample, this data contains a path to a photo in the Places365 dataset, a plausible user input related to the scene (User Input), a high-level plan that a robot could follow to address the user input (High_Level_Plan), and a short… See the full description on the dataset page: https://huggingface.co/datasets/jrfish/SelfReVision.
