This subdir contains certain subsets of the full set of images in this dataset.

The purpose here is to provide a way to easily download the images used to create a particular AI model.
So for example, to grab the images used in the "onegirl200" model, first grab
onegirl200.txt
Then use either "hf-download.sh <onegirl200.txt" to download from here  (huggingface.co)...
or use "dan-download.sh <onegirl200.txt" to download directly from the original site, danbooru.donmai.us

The tradeoffs in choosing each download script are:

* hf-download.sh will grab all the files, but will potentially download more than you need.
  If you only need one file from data-0000-cleaned.zip, it will have to download and extract the entire file.
  As long as you have at least 10Gig of free space, this shouldnt be a big deal

  It will extract the files in the same format the zipfiles have them in

  I personally would use hf-download.sh


  

* dan-download.sh will download only the files that you need. However, you will have to create a
  login and an "API key" (for free) for the danbooru site, and then edit the script to include them,
  before you can make use of it

  Additionally, it will download all files to the CURRENT directory, whereas hf-download.sh will
  keep the files in the same subdirectory format they are here

  Also, it is quite slow. Perhaps 1 file per second