Master Generative AI : Ultimate Cheat Sheet

Install diffusion AUTOMATIC1111 on windows by setting up python and conda, git, and optional visual studio, clone the repo, place models and gifsgan, and run webui-user.bat to generate AI art.

Explore the Automatic1111 txt2img workspace, configuring models, prompts, sampling methods, steps, and acceleration options. Learn to use extensions, scripts, extra networks, and image to image, inpainting, and upscaling tools.

Master Automatic1111 workspace img2img to generate new images from a base image, using prompts, sampling, and tools like inpaint, sketch, and aspect ratio controls.

Navigate the automatic1111 workspace to upscale images losslessly, using resize and multiple upscalers, then batch process with input/output directories while applying face restoration and background removal.

Explore automatic1111 checkpoint merger to integrate up to three models A, B, and C, adjust weights with multipliers, and create a merged model in .ckpt or .safetensors formats.

Train custom embeddings or hypernetworks in stable diffusion using textual inversion. Prepare pre-processed training images with captions, adjust learning rate and prompts, and save the model with a unique name.

Explore the stable diffusion 1.6 update with lycoris features merged into the native lora extension, plus metadata and trigger word management and new upscalers for better image generation.

Learn to build a basic prompt by combining the raw prompt with a chosen style, action/scene, and artist to shape the final image using filters and resolution.

Explore using negative prompts to control image output by specifying what to exclude, with examples, while acknowledging imperfect results as you refine.

Explore community prompts and templates to create ai art with stable diffusion. Learn to adjust prompts, styles, sampling steps, cfg scale, and lora models for varied, detailed results.

Explore how to use lora models to generate and refine images through community prompts and model variants. Learn how to customize prompts and seeds to improve final results.

Master generative AI introduces techniques like textual inversion, hypernetworks, Lora, and LyCORIS to generate landscapes and portraits through prompts, with experimentation and fine-tuning for surprising results.

Learn to fix eyes with inpaint by masking, upscaling, and inpainting masked areas, adjust sampling steps, restore faces, and control noising strength to generate improved eyes.

Install the depth library from a URL and enable control net to fix deformed hands by adjusting background, width, height, depth, and saving hand maps as png.

Explore image-to-image generation by adjusting the base image, width, height, and aspect ratio; experiment with denoising strength from 0 to 1, sampling method euler a, and prompts for varied outputs.

Generate stunning art using inpaint by applying prompts, negative prompts, and mask-based edits with brush size and color controls to create new images.

Combine inpainting and sketch inpaint to regenerate only the painted area. Adjust resize mode, mask blur, mask transparency, paint area mode, masked padding, and denoising strength before you generate.

Perform style transfer with control net in sd webui, uploading content and style images, selecting depth preprocessors, and tuning denoising strength, control weight, and control steps.

Explore Stable Diffusion with the xyz plot to test parameters by adjusting x, y, and z types and values, enabling legends and seeds options to compare images and view sub-grids.

Generate unlimited prompts for AI art by using templates, adding artists and negative prompts, exploring prompt gen extensions, and refining outputs with sampling controls and wildcards.