🚩 Report: Illegal or restricted content
This dataset was scraped from Cara, a platform that strictly prohibits scraping and AI training off its art in its ToS.
This data is also poisoned with nightingale and other algorithms to make training data unusable. This was clearly not removed as CaptiveDreamer uses direct links to the original photos, meaning the “poisons” are still in tact.
As such, this dataset should be removed from the platform for the integrity of training data, legal training, and to respect the wishes of the owners of the artwork.
Nightshade and other "poisoners" are ineffective and can be simply removed.
Nightshade and other "poisoners" are ineffective and can be simply removed.
Not all of them are simply removed. OP is clearly not knowledgeable enough to bypass all of the above as they use direct links to the original photos, so these “poisons” are clearly still in tact.
Also, OP has still broken the ToS to obtain the data in the first place.
AI training should be conducted in a legal, ethical, and clean manner, not to laugh in the face of creators that have concerns.
https://smallbusiness.chron.com/urls-copyrighted-62319.html