Stable diffusion: Easy 4K AI generated images with Low VRAM GPU

Everything is explained in the video subtitles. Next video I’ll show you how to generate 8K images with way more detail, still with 8GB VRAM. This video show how to make a 4K image with only an 8GB VRAM. 1 - Draw a sketch(text2img): use low res images to generate an idea of what you want 2 - Generate new image from sketch: Send sketch you like to img2img then put the max resolution your GPU can survive(Depend on VRAM and launch options), use 0.5 denoise. 3- Inpaint things you don’t like and change them with the prompt (inpaint masked content option) 4 - Upscale to desired size (I like 4x img size) For step 2 you’ll have to find what your GPU can handle for its max image size(for me its 1248 x 816 with launch options) or 800x800 with no launch options. My launch parameters: --no-half --always-batch-cond-uncond --opt-sub-quad-attention --medvram --disable-nan-check List of launch parameters: