{
  "description": "[b][u]How It's Made:[/u][/b]\n* I generated an initial image using txt2img. The parameters used are in the attached text file and the image generated is included as the last file.\n* If there's parts of the original generation I'm not happy with I'll inpaint those pieces until I'm happy with the general vibe of the image.\n* I run that through the upscaler at 2x to get a higher resolution image\n* Then I inpaint individual pieces of the image to get more details. The prompt is adjusted for each inpaint. For example I may select a paw and the prompt would only include keywords related to paws and fur.\n* Once I'm happy with the full image (Almost every piece of the image gets inpainted over) I may upscale it 2x again.\n* If anything got messed up by the upscaling I'd do another targeted inpainting pass until I'm happy with the final image.\n* Sometimes I'll pull a final image into Photoshop for some color correction.\n\nI use InvokeAI for all of the generation and inpainting."
}