search

IDEA-Research / HumanSD

[ICCV 2023] The official implementation of paper "HumanSD: A Native Skeleton-Guided Diffusion Model for Human Image Generation"
Apache License 2.0
271 stars 18 forks source link

LAION-Human Dataset #29

Open unrealMJ opened 10 months ago

unrealMJ commented 10 months ago

Hi,

I have already downloaded the full laion-5b dataset. How can i use your .parquet and mapping file to get corresponding image.

unrealMJ commented 10 months ago

Also, the .parquet has 2.86M images, while the mapping.json has 1M images, it seems that is a subset of .parquet. I'd like to ask for the details about .parquet, i think is a subset of laion-5b, how do you get it?

juxuan27 commented 10 months ago

Hi, @unrealMJ ! Thank you for your focus. You may use python utils/download_data.py to download all images. The .parquet has provides images in Laion-Aesthetic since we have a different order with the original Laion-Aesthetic dataset as mentioned in issue4.

unrealMJ commented 10 months ago

Hi, thanks for your reply. The Laion2b-en-aesthetic in huggingface has 52.1M rows, but the .parquet you provided only has 2.86M rows, i'd like to ask the difference.

juxuan27 commented 8 months ago

The .parquet we provide is a subset of Laion2b-en-aesthetic, filtering out the part with a higher aesthetic score.