In 2023, the birth of AI painting.
Let me start by sharing my views on AI painting. As a former digital art migrant from an art major, I personally have no objection to AI painting. As a tool, it is truly convenient, time-saving, and cost-effective, saving me time on production and reducing the need for outsourcing.
Unfortunately, for most art practitioners, we are currently in a time of transformation. This transformation will eliminate most artists who love art but lack sufficient skills to make a living from it, leaving only the most exceptional artists. This is good for art, but not so good for most practitioners. Just like how AI painting has eliminated traditional painting from a commercial perspective, from the perspective of an observer, AI has in essence "bled" from the Chinese painting migrant workers during its initial training. Now, it is reversing the process and replacing them. After the emergence of AI, many outsourcing companies in the art industry have begun layoffs and replaced them with AI. Combined with the current employment environment, this is why many practitioners in China hate AI.
However, consider this: isn't every technological advancement in history been淘汰to previous generations of workers? Just like how the invention of combines eliminated professional wheat harvesters, each small step of progress in this age overwhelms individuals. But the wheels of progress crush those who cannot keep up, and we must adapt to the changing times. So, young friends, when the combines are already at our doorstep, do not complain. Lay down your sickles and become the first drivers of these combines!
===========================================================
The following is a compilation of 154 selected exam sketches.
Trigger words: quick sketch, greyscale, monochrome, simple background, traditional media, white background. Do not use prompts that conflict with desired effects.
The base template is ChilloutMix. I have tested it and found that it works well in most excellent realistic models. Anime and 2.5D models can also be used, but they deviate from the original intention of the title "Joint Exam Sketching". Please use it carefully.
It is recommended to use tools like "ControlNet" and "OpenPose" to control actions for better results.
The recommended number of iterations is between 20 to 40 steps, depending on the situation.
The recommended sampling method is DPM series or Euler, Euler a, restart.
The recommended size is between 512 to 1024 pixels.
The recommended prompt coefficient (CFG Scale) is between 7 to 12.
The LoRA weight varies depending on the situation and generally floats around 1.
For high-resolution repair (Hires. fix), the re-drawing amplitude can be between 0.3 to 0.5 to ensure maximum image structure, or between 0.5 to 0.7 for more freedom! It's up to you!
If needed, you can turn on After Detailer.
===========================================================
Other works in this series:
"Art Students' Tears: Joint Exam Sketch | Style LoRA"
https://civitai.com/models/267797/the-tears-of-artists-joint-exam-sketch-portrait-or-style-lora
"Art Students' Tears: Joint Exam Coloring | Style LoRA"
"Art Students' Tears: Joint Exam Sketch | Style LoRA"
https://civitai.com/models/267852/art-students-tears-joint-exam-sketch-or-style-lora