Reverted the merge that was added to the HB1.5 finetuned datasets and tweaked the model with a new merge added in. The previous merge caused issues and now it seems to be producing better results.

Use Clip Skip 1 for best results

Needs VAE https://huggingface.co/stabilityai/sd-vae-ft-mse-original/tree/main

Instructions (thanks to XTM from our discord mod group) Click this to download the Stable Diffusion 1.5 VAE: https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt Save it in \stable-diffusion-webui\models\VAE\ Go to Settings Tab > User Interface and paste this in the Quicksettings List field: sd_model_checkpoint, sd_vae, sd_hypernetwork, CLIP_stop_at_last_layers, sd_hypernetwork_strength Click Apply Settings at the top of the page, then reload the WebUI (This doesn't mean to only refresh the page, but Reload the UI, OR close the CMD window for the WebUI and reopen it.) Once the WebUI has finished loading in command prompt, refresh the WebUI page, and you should have new settings at the top of the screen. Click SD VAE, then select vae-ft-mse-840000-ema-pruned. Leave it on unless you switch to an Anime model.

Uncheck your setting: Ignore selected VAE for stable diffusion checkpoints that have their own .vae.pt next to them

• vagina closeups, various poses of women with vagina's spread 

• bdsm rope data etc 

• upskirt photos, low angle butt/ vagina shots 

• blowjob images 

• various intercourse positions

• Male nude selfies