add violence detection prompts

Thanks to https://github.com/ml-research/OffImgDetectionCLIP usage: first embedding is safe, second is unsafe compute similarity between your embeddings and these two, if the closest is the safe embedding then it's safe, otherwise it's unsafe these have been trained on https://osf.io/2rqad/ and are good at detecting violent content
LAION-AI · Apr 10, 2022 · c2f21fe · c2f21fe
1 parent 057de24
commit c2f21fe
Show file tree

Hide file tree

Showing 2 changed files with 0 additions and 0 deletions.
diff --git a/violence_detection_vit_b_32.npy b/violence_detection_vit_b_32.npy
diff --git a/violence_detection_vit_l_14.npy b/violence_detection_vit_l_14.npy