metadata

license: mit
task_categories:
  - tabular-classification
language:
  - en
tags:
  - art
pretty_name: diffusionDB Prompts - 1,8mil
size_categories:
  - 1M<n<10M

To get access, you'll need to provide compelling reason,

you can contact me at: [email protected]

Since this database is WIP for future GPT learning.

Dataset made out of subset: "large_all" from https://huggingface.co/datasets/poloclub/diffusiondb.

Filtered from:

duplicate lines
blank lines
lines with single word
mathbold and similiar Unicode
http/s links
unprintable/binary characters
NUL characters
multiple whitespaces together
special characters like: ( [ { @ # $ % < etc.

other small stuff