Tensor.Art
Create

Alma Photographic PDXL

CHECKPOINT
Original

Updated:

v2
v1

This model aims for SDXL-like photorealism with minimal loss of Pony nsfw and anatomy capabilities. Use Pony style prompting.

v2 has SDXL VAE baked in.

Recommended Parameters:

Sampling and Schedule: DPM++ 2M SDE SGM Uniform

Sampling Steps: 20-50

CFG Scale: 6-8

Highres Fix:

Upscaler: DAT-4x

Denoising Strength: 0.4

Hires steps: roughly half of sampling steps

Face Detailer use denoise 0.4 or less.

These are just parameters that I've liked with my own hardware and preferences, use them as a starting point if you don't know what to try but feel free to experiment with other parameters and find which works best for you!

This model is a MBW of the following models and LORAs:

v2

Pony Realism v2.2 Main+VAE

RealVisXL v5.0 (BakedVAE)

v1

Pony Realism v2.2 Main+VAE

ClarityXL v2.0

Hands v4 XL

Perfection SDXL v1.0

TgirlsXL1.5

I post updates about what I'm working on at https://ko-fi.com/almascar you can also support me there ;)

Version Detail

Pony
SDXL VAE is baked in. Steps: 20-50 Sampler and Scheduler: DPM++2M SDE and SGM Uniform CFG: 6-8 Use Highres Fix and Face Detailer as needed typically denoise 0.4 Many other settings will work above are just my preferences for style and gen time. This is a completely new merge. MBW Merge of RealVisXL v5.0 and Pony Realism v2.2 Main This one is closer to the target photographic style, much better at prompt adherence, and Pony concepts compared to v1. I post updates about what I'm working on at https://ko-fi.com/almascar you can also support me there ;)

Project Permissions

    Use Permissions

  • Use in TENSOR Online

  • As a online training base model on TENSOR

  • Use without crediting me

  • Share merges of this model

  • Use different permissions on merges

    Commercial Use

  • Sell generated contents

  • Use on generation services

  • Sell this model or merges

Related Posts