Datasets:
Tasks:
Image-to-Text
Sub-tasks:
image-captioning
Languages:
English
Multilinguality:
monolingual
Size Categories:
10M<n<100M
Language Creators:
found
Annotations Creators:
found
Source Datasets:
original
ArXiv:
License:
Is there any way to get an 'all' subset of this dataset?
#4
by
surya-narayanan
- opened
I want to train on the whole thing and not subset it one by one, is that possible?
Yes, the loading script of this dataset has implemented a configuration named "all":
ds = load_dataset("red_caps", "all", split="train")
albertvillanova
changed discussion status to
closed
Thanks so much!