This repo contains the full captions for the MSCOCO dataset.
To load the dataset on your local device:
Local loading
from datasets import load_dataset
dataset = load_dataset("/home/think3/Desktop/1. MSCOCO_captions_dataset_edited/dataset_test_jsonl/ImageCaptions.py", split='train[{}:]'.format(1),cache_dir="dataset_test_jsonl/caching")
To load the dataset from Huggingface:
Test the remote repo:
from datasets import load_dataset
dataset = load_dataset("Arabic-Clip/mscoco_jsonl_full"… See the full description on the dataset page:
https://huggingface.co/datasets/Arabic-Clip-Archive/mscoco_jsonl_full.
