Spaces:
Running on Zero
NSFW Content Flagged regardless of its nature
Even text prompts/images that contain adulterated material which is completely normal and accepted by rules of nature/society are flagged by the models. CONTENT POLICY restricts nsfw generations involving minors, harassment, non consensual acts aka R@p3; NOWHERE it is explicitly mentioned that ALL and EVERY nature of nsfw content generations are restricted altogether. So why the flag on a grown man and a woman trying to get frisky ?!?!
The message box states to try safe prompting, but no matter how sugarcoated/civil the language is put, generations are flagged regardless. WORST PART is that the message is displayed AFTER the required Zero-GPU quota is eaten up. Honest Practices dictate that the programming of the models should indicate user "errors" (intentional or otherwise) PRIOR to starting the generation process, many third party generators outside of huggingface does that before wasting user credits.
Screenshots provided regarding the revised CONTENT POLICY for the sake of transparency.
Now it is over