Skip to content

Dataset · Dongping-Li

Robotics, RL, and VQA training dataset

Dongping-Li/EMMOE-100

An English dataset of 10K–100K examples featuring scene descriptions, trajectories, and VQA tasks for embodied AI.

What's inside

An English-language training split containing between 10,000 and 100,000 entries combining images and text data. Based on the file structure, each episode contains scene descriptions (scene.json), key paths (keypath.json), and multiple text instruction options or descriptions (info.txt).

What you can do with it

  • Training and evaluating multimodal models (VLMs) for navigation and action planning in an environment
  • Training reinforcement learning (RL) agents on tasks with visual context and trajectories
  • SFT fine-tuning of visual question answering (VQA) systems for robotics
Tasks
Visual question answering, Reinforcement learning, Robotics, Question answering
Languages
English
Size
10K–100K rows
Downloads this month
25.7K
Likes
1

How to load it

from datasets import load_dataset

ds = load_dataset("Dongping-Li/EMMOE-100")