Midjourney has high fees and cannot pay, and the generation of desired pictures is limited and the style is fixed. If you can quickly build your own AI painting system in 10 minutes, and there is no requirement for computer configuration, would you be tempted?
Let’s take a look at the screen after the AI system is built. You only need to enter a custom prompt to generate the desired portraits, landscapes, animals, various subjects, sci-fi, oil paintings, and realistic pictures. You can also use them later. Overlay super-resolution technology (SR) to further generate high-definition large images! It's hard to imagine that this is a wheel that can be made in 10 minutes. If you think the picture is familiar, yes, this is the famous stable-diffusion open source diffusion model!
Not much to say next, just follow the fool-like operation steps one by one to complete the AI painting system construction! First, the first step is to open the corresponding github link: https://github.com/camenduru/stable-diffusion-webui-colab, which is open-sourced by camenduru and uses google-colab resources to perform stable-diffusion diffusion model fine-tune. List of out-of-the-box models.
Immediately after the second step, ctrl+F or command+F search for ChilloutMix, and click on the stable version to enter the next step.
The third step is to jump to the Google Colab interface through the link, which is a cloud machine that provides GPU resources for free (if you do not have a Google account, you need to register yourself).
Then start the fourth step, click the run button in the upper left corner of the picture above to run the code in the box, ignore the warning and click still to run, wait for a few minutes, during which it will start to download the model, prepare the webUI and print a series of logs, just take a look.
Following the fifth step, confirm that the fourth step is carried out normally, and produce 4 links at the end of the log. These links actually correspond to the remote interface of the webUI, just click the last link.
The sixth and final step is to enter the stable-diffusion web interface built by yourself, enter a custom prompt in the first translucent gray box, and write a Negative prompt (optional) in the second gray box, respectively Represents forward prompts and reverse prompts, further adjust the Sampling method to DPM++ SDE Karras, set the Sampling steps to 20-40, Width and Height represent the width and height can be set freely, then click Generate and wait for the drawing (1min for the first time ).
Of course, writing prompt is also a skill, perhaps the most important thing to be able to produce the pictures you want. You can enter https://prompthero.com/ website to refer to the essence of prompt experience from the vast number of netizens around the world to write your own prompt.
A series of pictures and prompts are attached to the following picture for reference only~
Prompt1:professional portrait photograph of a gorgeous Norwegian girl in winter clothing with long wavy blonde hair, ((sultry flirty look)), freckles, beautiful symmetrical face, cute natural makeup, ((standing outside in snowy city street)), stunning modern urban upscale environment, ultra realistic, concept art, elegant, highly detailed, intricate, sharp focus, depth of field, f/1. 8, 85mm, medium shot, mid shot, (centered image composition), (professionally color graded), ((bright soft diffused light)), volumetric fog, trending on instagram, trending on tumblr, hdr 4k, 8k
Negative prompt:(bonnet), (hat), (beanie), cap, (((wide shot))), (cropped head), bad framing, out of frame, deformed, cripple, old, fat, ugly, poor, missing arm, additional arms, additional legs, additional head, additional face, multiple people, group of people, dyed hair, black and white, grayscale
Prompt2:Hyper realistic painting of a beautiful girl in an EVA plugsuit, hyper detailed, anime, by greg rutkowski, trending on artstation
Prompt3:THE CHERRY BLOSSOM TREE HOUSE :: beautiful ornate treehouse in a gigantic pink cherry blossom tree :: on a high blue grey and brown cliff with light snow and pink cherry blossom trees :: Roger Deakins and Moebius and Alphonse Much and Guweiz :: Intricate details, very realistic, cinematic lighting, volumetric lighting, photographic, --ar 9:20 --no blur bokeh defocus dof --s 4000
Prompt4:bunny, high ornamented light armor, fluffy fur, foggy, wet, stormy, 70mm, cinematic, highly detailed
In addition, this system not only supports txt2img (Wensheng diagram), but also supports img2img (image generation diagram), Train (fine-tune fine-tuning training), etc. In addition, this method will be limited by Google bandwidth and resources, and it can also be considered on a PC stand-alone (Windows&Mac) Configure your own AI painting system, but it is a little more complicated, and will be updated later~
It's not easy to be original, and watching collections is a great motivation for sharing! If you have any questions, please leave a message directly~
History recommendation
Original super detailed Python learning route + mind map
The original and most complete data analyst learning route + mind map!
Data Analysis of "High-Frequency Face-to-face Classics"
"High-frequency face-to-face classics" machine learning
In-depth learning of "High-frequency face-to-face classics"
Selenium crawls Jingdong products
Scrapy framework crawls QQ music
Data Analysis and Mining
Data Structures and Algorithms
Deep Learning and Big Data Framework
Welcome to pay attention, thank you for "watching", it's rare to follow the fate~