Open X-Embodiment
1M+ real-robot trajectories from 20+ institutions.
RoboCerebra VideoQA Benchmark This directory contains the Video Question Answering benchmark for evaluating RoboCerebra models on action progress and…
RoboCerebra VideoQA Benchmark This directory contains the Video Question Answering benchmark for evaluating RoboCerebra models on action progress and task completion understanding. Files eval.py: Evaluation template script. test.json: Test set for progress estimation. qwen25vl vlm data/: Training data.
Source: Hugging Face Hub (qiukingballball/RoboCerebra_Progress_Estimation). Metadata imported from the dataset’s Hub tags.