search

rom1504 / img2dataset

Easily turn large sets of image urls to an image dataset. Can download, resize and package 100M urls in 20h on one machine.
MIT License
3.62k stars 336 forks source link

laion-coco not available #372

Closed vanpersie32 closed 9 months ago

vanpersie32 commented 9 months ago

laion-coco not available. I cannot download the parquet files from hugging face official webdataset.

vanpersie32 commented 9 months ago

download the https://huggingface.co/datasets/visheratin/laion-coco-nllb/blob/main/data/test-00000-of-00001-dc1609aa34aab9fb.parquet instead