Use either wdv3 captions or real captions, i personally use Florence-2 Large in TagGUI
Train on base pony, results are close enough imho
Use Prodigy Optimizer with a d-coeff of 0.5
Use masked training with an 1.0 alpha mask for the face and 0.3 for the body, if possible remove hands and feet from the mask
Do not train the text-encoders, only unet
600 - 1000 steps is suggested

Suggested Settings:

Sampler: Euler a + AYS (Align your steps)
Resolution: 832x1216
Steps: 14
CFG: 4 - 7
Perturbed Attention Guidance 2.7

HiRes Upscaler: 4x_NMKD-Siax_200k
Sampler: DPM++ 2M SDE + SGM Uniform
Steps: 30 - 50 (Or 12 with SDXL Hyper LoRA at 0.5 weight)
Denoising Strength: 0.25 - 0.35

Prompt:

As usual start with the score tags. I personally start every prompt with this huge list of tags:

score_9, score_8_up, score_7_up, photo, source_photo, source_real, RAW, 8k, low quality, traditional media, amateur, real life photo, film grain, dark, darkness, dim night light, light flash, very dark focused flash photo, (dark background:0.75), low light, detailed face

DatAss Rev 2 Recipe:

Step1: GameArt Pony + CyberRealistic Pony (0,0.1,0.2,0.3,0.3,0.3,0.3,0.3,0.3,0.2,0.2,0.2,0.3,0.3,0.3,0.3,0.3,0.3,0.2,0.1)
Step2: Step1 + Custom Trained Pony Model (0,0.15,0.25,0.3,0.25,0.2,0.15,0.15,0.1,0.1,0,0.1,0.1,0.15,0.15,0.2,0.25,0.3,0.25,0.15)
Step3: Step2 + (Damn Pony Realism - CyberRealistic) (0,0.35,0.6,0.65,0.45,0.25,0.1,0.1,0.05,0.05,0,0.05,0.05,0.1,0.1,0.25,0.45,0.65,0.6,0.35)
Step4: Step3 + (Personal Anteros Mix - SDXL) (0,0.25,0.4,0.45,0.25,0.15,0.1,0,0,0,0,0,0.05,0,0.1,0.15,0.25,0.45,0.4,0.25)
Step5: Step4 + (Step4 (+SG Finetune) - Step4) (0,0.2,0.55,0.43,0.35,0.1,0.05,0.05,0.05,0.05,0.05,0.05,0.05,0,0.05,0.05,0.05,0.05,0.05,0.05,0.05,0.1,0.35,0.43,0.55,0.2)
Step6: Step5 + Personal Anteros Mix (0,0.1,0.28,0.28,0.2,0.1,0,0,0,0,0,0,0,0,0,0.1,0.2,0.28,0.28,0.1)
Step 7: Step6 + ICKPOTLycoXL 0.25 + photo_v2 0.2