->SHOWCASE HAVEALL MODEL<- is featured on the model's website. The site shows several hundred examples with images and complete prompts for generation.

The model is trained for over 20,000 images at 1024x1024 resolution.

Recommendations for generating images with the Haveall model:

- Standard image resolution for generation: 768х768

- You can generate images in other formats: 768x1024, 1024x1024, 768x960.

- No Lora and embeddings are required, just simple prompts.

- For best results use the VAE, vae-ft-mse-840000-ema-pruned https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt

- 4x-UniScale-Balanced [72000g].pth

I do it this way: I generate a test image in 768x768 format, if I am satisfied with the result of the composition, I set the same seed, leave the resolution at 768x768, check Hires. Fix and Upscale by 2. And generate the image again, the image is obtained with a resolution of 1536x1536 with very good quality and detail. As an upscaler I use 4x-UniScale-Balanced [72000g].pth and DPM ++ 2 SDE Karras sampler.