LocalLLaMA

Looking for smallest VLM for NSFW image detector (atleast 5 it/s on CPU)

Hello everyone, I am looking for a very small VLM or Transformer based ViT, which will inference over images (each size less than 10MB, any ratio/resolution possible). The model should return 1 or 0 that the img is NSFW or not, thats it. I want the mod…