Resized to 41% of original (view original)
| Prompt | score_9, score_8_up, score_7_up, score_6_up, 3girls sitting side by side, beach, outdoors, bikini, large breasts, smile, looking at viewer, ADDCOMM, mamarissa, black hair, braid, hair over one eye, colored inner hair, red hair, red eyes, horns, ADDCOL, pekomama, hololive, mature female, rabbit girl, blue hair, single braid, hair over shoulder, orange eye, ADDCOL, Rinmama, short hair, grey hair, ahoge, hair between eyes, green eyes, earrings, pearl necklace, jewelry, <lora:TruckKun_Style_Lora_V3:1>, <lora:mamarissa-pony:0.4>, <lora:Pekomama:0.4>, <lora:Rinmama-step00002000:0.4> |
|---|---|
| Negative prompt | aidxlv05_neg, score_1, score_2, unaestheticXL2v10, japanese text, english text, text, disembodied penis |
| Sampler | Euler a |
| Seed | 3290954315 |
| Steps | 25 |
| Cfg Scale | 5 |
| Model Hash | d3ee23d452 |
| Denoising Strength | 0.4 |
| Model | autismmix_DPO |
| Rp Active | True |
| Rp Calc Mode | Attention |
| Rp Divide Mode | Matrix |
| Rp Mask Submode | Mask |
| Rp Matrix Submode | Columns |
| Rp Prompt Submode | Prompt |
| Rp Ratios | 1,1,1, RP Base Ratios: 0.2, RP Use Base: False, RP Use Common: False, RP Use Ncommon: False, RP Options: [False], RP LoRA Neg Te Ratios: 0, RP LoRA Neg U Ratios: 0, RP threshold: 0.4, RP LoRA Stop Step: 0, RP LoRA Hires Stop Step: 0, RP Flip: False, Lora hashes: TruckKun_Style_Lora_V3: 9f833fe68eb4, mamarissa-pony: 1468749eb6fc, Pekomama: dabcc08b4806, Rinmama-step00002000: 85d1e6543630, TI hashes: aidxlv05_neg: 84a6266ecaa6, unaestheticXL2v10: c792360b80a1, aidxlv05_neg: 84a6266ecaa6, unaestheticXL2v10: c792360b80a1 |
| Size | 2067x1414 |
| VAE | sdxl_vae.safetensors |
| VAE Hash | 235745af8d |
| Version | f0.0.17v1.8.0rc-latest-276-g29be1da7 |
- ‹ prev Search: devil_horns next ›
Anonymous 04/09/24(Tue)11:38:50 No.73309404
1687683682929519.png (3.44 MB, 2064x1408)
>>73308018
Tough with multiple LoRA since it depends on how well they were trained and just general degradation with multi-LoRA. If they are in the model already I would use couple and then upscale with RP ADDCOMM,.
If multi-lora I would use RP for base gen using ADDCOMM, fix base gen with inpaint and the upscale.
Example
Base gen
https://files.catbox.moe/3lwb3u.png
Fixed, inpaint
https://files.catbox.moe/b0w4yd.png
Lower LoRA weights on upscale (except style LoRA, so everybody was 0.4 here)
Final
https://files.catbox.moe/wpl85q.png
Another way is gen 3 close to feature characters and then inpaint their faces, probably will cause the least hit to quality.
>>73308466
Alternate universe Fuwamoco! XD
>>73308965
SEXO, Love Tomoe
>>73309114
Good morning! Yeah love the eyes on this style. How did the new Clara go?