Charts look simple, but answering a question about one can require recovering values, understanding visual structure, and connecting both to language. ChartNet’s key contribution is to align these different representations, making it possible to train and assess systems on more than image captioning alone.
AI Dataset2026
ChartNet: A Million-Scale Multimodal Dataset for Chart Understanding
Aligns chart images with plotting code, underlying tables, summaries, and reasoning-based QA for chart extraction and interpretation. Specialized subsets add human-verified, multi-panel, grounded, and multilingual examples.
Introduction
Information
- Websitehuggingface.co
- OrganizationsMIT, MIT-IBM Watson AI Lab, IBM Research, Abaka AI, 2077AI
- AuthorsJovana Kondic, Pengyuan Li, Dhiraj Joshi, Isaac Sanchez, Ben Wiesel, Shafiq Abedin, Amit Alfassy, Eli Schwartz, Daniel Caraballo, Yagmur Gizem Cinar …
- Published date2026/03/31
Categories
More Items
AI Dataset2026
Provides the three-stage training data for Carbon-X, a DNA language model: large-scale genomic and biology-text data for mid-training, plus supervised fine-tuning and reinforcement-learning examples.
AI Dataset2026
Provides executable programs for generating 3D content, paired with captions, renders, and available verification records. It spans multiple 3D coding environments, with source-specific licenses and varying sample quality.
AI Dataset2026
Provides source-grounded, long-form instruction data for defensive cybersecurity assistants across 17 domains. Mixed-source licensing and reusable reference appendices require care when training on or redistributing the data.