Dataset · HuggingFaceFV
Videos with descriptions and questions
HuggingFaceFV/finevideo
An English video corpus for training models to answer questions about videos.
What's inside
The dataset contains video clips and text data for visual question answering and video-text-to-text tasks. The language is English, and the size ranges from 10k to 100k examples. The card describes the structure, loading process, and CC-BY license; the source and collection method are not disclosed in the available snippet.
What you can do with it
- Fine-tune a video-language model or LoRA for video VQA
- Evaluate a video-text-to-text system on English clips
- Tasks
- Visual question answering, Video text → text
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 42K
- Likes
- 411
- Access
- on request on Hugging Face
Models trained on it
- HuggingFaceTB/SmolVLM2-500M-Video-InstructImage text → text 1.7M
- HuggingFaceTB/SmolVLM2-2.2B-InstructImage text → text 194.6K
- HuggingFaceTB/SmolVLM2-256M-Video-InstructImage text → text 44.7K
- OpenGVLab/InternVL2_5-8BImage text → text 18K
- OpenGVLab/InternVL2_5-4BImage text → text 17.9K
- OpenGVLab/InternVL2_5-2BImage text → text 14.5K
- OpenGVLab/InternVL2_5-1BImage text → text 7.6K
- OpenGVLab/InternVL2_5-38BImage text → text 4.8K
How to load it
from datasets import load_dataset
ds = load_dataset("HuggingFaceFV/finevideo")