---
license: creativeml-openrail-m
tags:
- stable-diffusion
- text-to-image
---
---

# P4A Mandala

trained with kohya_ss (edg settings)

pinkal09 dataset deviantart.com/pinkal09

trigger "p4a psychedelic" + EasyNegative

huggingface.co/LibreSD/Various/resolve/main/EasyNegative.safetensors

huggingface.co/LibreSD/Various/resolve/main/EasyNegativeV2.safetensors

samples civitai.com/models/147191/p4a-mandala

---

# Merge Info

step1_hard = [p4a-step2](https://huggingface.co/Powidl43/psychedelic/tree/main/step2-merge) 0.6 + pinkal09 0.4

step2_hard_a = step1_hard-camelliamix_v3 0.6 + step1_hard-greymix_v2 0.4

step2_hard_b = step1_hard-counterfeit_v3 0.6 + step1_hard-nabimix_v2 0.4

p4a_mandala_v1_hard_a = step2_hard_a 0.6 + step2_hard_b 0.4

p4a_mandala_v1_hard_b = step2_hard_b 0.6 + step2_hard_a 0.4

step1_soft = [p4a-step2](https://huggingface.co/Powidl43/psychedelic/tree/main/step2-merge) 0.8 + pinkal09 0.2

step2_soft_a = step1_soft-camelliamix_v3 0.6 + step1_soft-greymix_v2 0.4

step2_soft_b = step1_soft-counterfeit_v3 0.6 + step1_soft-nabimix_v2 0.4

p4a_mandala_v1_soft_a = step2_soft_a 0.6 + step2_soft_b 0.4

p4a_mandala_v1_soft_b = step2_soft_b 0.6 + step2_soft_a 0.4

---

base models and other essentials huggingface.co/LibreSD