Realism from HaDeS' PoNy V1

This model was created by request of some followers. It is based on my checkpoint Realism from HaDeS XL with LoRa's trained with Pony and adding some LoRa from my checkpoints Anime from HaDeS XL and R.O.A.R. XL.

For a V1, I tried to add as much versatility as possible, although I warn in advance that the checkpoint has a lot of room for improvement. I will be updating in the next weeks and you can expect a new improved version every 2-3 weeks starting in September. (I will have vacations and more time to train, mix and test).

Recommended Settings:

• Width/Height: 896x1152
• Sampler: Euler A
• Steps: around 30
• CFG: around 4-7
• VAE: Already baked in.
• Highly recommended to add to the prompt: score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up,
• Negative Prompt: score_1, score_2, score_3.
• HiRes: 4x_NMKD-Siax_200k with half the regular steps and 0.15 Denoise + x2 Upscale