nvidia/Nemotron-Image-Training-v3
Visual Question AnsweringImage Text To TextEnglish
Created by nvidia at 2026, the nvidia/Nemotron-Image-Training-v3 is a visual question answering dataset in English in Parquet format.
About nvidia/Nemotron-Image-Training-v3
Nemotron Image Training v3
Versions
Date
Commit
Changes
2026-04-28
HEAD
Initial commit.
Dataset Description
Nemotron Image Training v3 is a collection of image-centric multimodal training data for vision–la...
Details
- Task
- Visual Question Answering, Image Text To Text
- Language
- English
- Format
- Parquet
- Rows / instances
- N/A
- Creator
- nvidia
- Year
- 2026