Datasets:
Modalities:
Image
Formats:
imagefolder
Languages:
English
Size:
10K - 100K
ArXiv:
Tags:
computer-vision
vision-language
visual-question-answering
image-segmentation
object-detection
image-classification
License:
Download data/Distress_PaveVQA/correction.json from WLL123/PaveBench: direct link, hf CLI and curl.
- Browser
- Download file 863 kB
-
https://huggingface.co/datasets/WLL123/PaveBench/resolve/main/data/Distress_PaveVQA/correction.json
- Command line
-
hf download hf://datasets/WLL123/PaveBench/data/Distress_PaveVQA/correction.json
-
curl -L -o correction.json https://huggingface.co/datasets/WLL123/PaveBench/resolve/main/data/Distress_PaveVQA/correction.json
863 kB
File too large to display, you can check the raw version instead.