Type | |
Stats | 500 0 |
Reviews | (47) |
Published | Oct 9, 2024 |
Base Model | |
Usage Tips | Clip Skip: 2 |
Hash | AutoV2 7D0748622D |
Welcome to the Magical World
when your fantasy & imagination becomes reality with AI
Information about this model:
This model is a trained model and a fusion model. trained with more than 1000+ datasets as preliminary, and merge them with some selected models, and tuned with the MBW technique to enhance the realistic and detailed nature of the subject without adding LoRA detail.
In the future, I hope I can create more,upgrade and optimize it more and move on to working on the SDXL model.
Hope you like, be creative and have fun!
V1.0
Useful Parameter Settings:
Recommended Sampler: DPM++ 2M Karras,DPM++ 2M SDE Exponential,& DPM++ 3M SDE
Step: 20 - 40
CFG: 4 - 8
HireFix upscaled by 2 with: 8x_NMKD-Superscale_150000_G ,4x_NMKD-Superscale-SP_178000_G,ESRGAN_4x or R-ESRGAN 4x+.
Hires step: 20
Negative Embeddings: badhandv4, ng_deepnegative_v1_75t,((BadDream, AS-Young-neg ,epiCNegative, By bad artist -neg)(optional)
VAE is a must: vae-ft-mse-840000-ema-pruned.ckpt
Extension: Adetailer, sd-dynamic-thresholding
V2.0 (Skipped)
V3.0 (Hyperrealistic looks, new faces, anatomy details upgrade)
V3.0 vs V1.0 Images result
Comparison:
DONT FORGET TO LEAVE A REVIEW.
サンキュー!