anime-rating-int8

Dynamic int8 (per-channel) quantisation of deepghs/anime_rating caformer_s36_v1_pruned_ls0.1: rates an illustration safe / r15 / r18. Input input NCHW 1x3x384x384 RGB in [-1, 1]; output softmax probabilities. 39 MB, probabilities within 0.02 of the fp32 model.

Used by Latent Studio as a second opinion beside a photo-trained NSFW classifier for images from anime models, which the photo classifier misreads as explicit.

License

MIT, as the original (DeepGHS).

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for latentdivergence/anime-rating-int8

Quantized
(1)
this model