Best for 2D Art. All example images are generated using just plain txt2img with no ADetailer / hiresfix.Please share your pictures and prompts so I can improve the next version. Recommended VAE can be ignored, I don't know how to remove it.Best settings for resolution:896 x 11521152 x 7681024 x 1024CFG:4-6Samplers:RestartDPM++ 2s A KarrasDPM++ 3M SDE ExponentialADetailer for person / face is recommended, 2x hiresfix with 4x_BooruGan_650k"score_9, score_8_up, score_7_up, score_6_up, score_5_up, source_anime" in positive prompt"photo,hyperrealistic,realistic, 3d" in negative prompt.Used some example character / style loras at the end of the sample pictures.thanks to @Consceleratus (Atilesscence) @insertusername (VividPDXL) @Anzhc (AAAAutism) @DankGabrillo (UncannyAF.PonyXL) for the models merged into @PurpleSmartAI (PonyDiffusionV6). I added my own custom trained lora into the mix.
