Uploaded the remastered version of OldDream v2.2, which added TCD and a bit of lightning, the model can now run from 8 steps, but is best left at 13 steps.
Some notes when using the TCD model:
Sampling method: Euler a (You can try other sampling methods, but most of them are unusable)
Sampling steps: 13 (If it is lower, it will lack important details, but if it is higher, redundant details will appear)
CFG Scale: 1,5 (Lower will not have enough accuracy or sharpness or vivid colors in the image, and anything higher will cause spots and color errors to appear)
Negative prompt: (This doesn't have to be the same as mine, perhaps yours might be even better.)
(worst quality:1.5), (low quality:1.5), (normal quality:1.5), lowres, bad anatomy, bad hands, multiple eyebrow, (cropped), extra limb, missing limbs, deformed hands, long neck, long body, (bad hands), signature, username, artist name, conjoined fingers, deformed fingers, ugly eyes, imperfect eyes, skewed eyes, unnatural face, unnatural body, error, painting by bad-artist
An improved version but still trying to keep the uniqueness of the previous v1 version. It is sharper, no more blobs, adds more diverse new data, adds more semi-realistic images, more beautiful faces, reduces too dense details to create a smoother feeling for the photo
If you like my content, support me at:
Patreon in here:
https://linktr.ee/Duc_Haiten
If you want to use my models to create unlimited images at a small cost without having to install anything, mage.space will be a good choice, i have some models that are exclusively available on there.
Negative prompt:
(worst quality:1.5), (low quality:1.5), (normal quality:1.5), lowres, bad anatomy, bad hands, multiple eyebrow, (cropped), extra limb, missing limbs, deformed hands, long neck, long body, (bad hands), signature, username, artist name, conjoined fingers, deformed fingers, ugly eyes, imperfect eyes, skewed eyes, unnatural face, unnatural body, error, painting by bad-artist
Steps: 30
Sampler: DPM++ 2M Karras
CFG scale: 7-9
CFG 7 will avoid misplaced or redundant detail errors, but higher CFG is better for images that need more detail added. Highest possible to 11
Description
FAQ
Comments (11)
Yea, loving the generations so far, however, I'm using all of your settings and my images are still coming out blurry. What resolutions do you recommend? Clip skip 1 or 2? Which VAE?
You need to send me detailed information about the cpu gpu, the type of webui and version you are using, and the system parameters installed on that webui.
intel i9 12900k, Nvidia RTX 3080 12GB, DDR5 64GB, Forge UI 17v1.8.0rc-latest-276
@9ball There are no problems with the hardware and webui, maybe it's the parameter settings when creating the image, send me those parameters
@DucHaiten Steps: 13, Sampler: Euler a, CFG scale: 1.5, Size: 1024x1024, Model hash: fa7c2f490d, Model: duchaitenOlddream_v22TCDLightning, VAE hash: 235745af8d, VAE: sdxl_vae.safetensors, Version: f0.0.17v1.8.0rc-latest-276-g29be1da7
I don't use extra 'vae', try removing vae, if it still doesn't work, try turning on hires fix with the 4x ultrasharp model
@DucHaiten What resolution did you use in the sample images?
@9ball 4x-UltraSharp
@DucHaiten Image dimensions?
@9ball 704x1024, In addition, I heard that there is a problem with adding certain loras to the lightning model and there is a possibility of image blur errors. Not all loras suffer from it, but it is possible
Easily one of the most legendary creators on this site. Never miss.