![]() If you are to color it digitally, a transparent background is more useful. If you are to color in analog, it is recommended to use a white background. This option is also useful to use as a coloring book with hints on where to paint in what color if you want to repaint with the same color as the original image. It's good for painting watercolor with a soft touch, and it's suitable for converting face photos without losing their original atmosphere or converting food photos without losing their delicious look. If you want to make the contour lines less noticeable when applying color, this mode is the best. The colors used in the original image will be extracted as outlines. This option is recommended for technical drawings such as drafting of buildings or machinery.Īlso, if you convert with "Fine Pen" and the lines are whited out or blacked out, try again with "Pencil" and it may work. Since the position of the lines is clear, it's suitable for tracing lines by yourself over the output image. Not all printers are supported, so this is a setting for advanced users. However, the vector format cannot be saved on iPhone or iPad. ![]() So it is recommended for larger images or printing. The difference in this vector format is that the output is not blocky(pixelated) no matter how much you zoom in. It looks almost the same as the regular "Smooth Pen". "Smooth Pen" style can be output in vector format (SVG file). It's recommended that the original image is 1280px or larger and is not too compressed. On the other hand, it's not good for converting illustrations which require a delicate touch, noisy photos, low-resolution images, and portrait photographs. This option is suitable for converting images with clear outlines, such as manga and anime, or flat color illustrations. Since the lines are clear, anti-aliased and have no shading, the output is similar to a coloring page for children, and it's especially suitable for printing. Like a fountain pen, the lines are pressure sensitive. It's like a G-pen (thick dipping pen) for drawing illustrations. When it comes to coloring books, adult coloring books are better suited than those for children. This can extract lines accurately and is suitable for fine-touch drawing.īecause of its wide range of usages and short processing time, it is recommended to choose this option usually. I will use 2.1 model (512-pixel) for this test.It's like a mapping pen (fine dipping pen) for drawing illustrations. Now let’s find a good universal negative prompt. We have already touched on the importance of negative prompts in v2. Excluding NSFW images also unintentionally biases the data towards the bad and ugly. It could be a failure of the filter, or its just be the nature of the NSFW images. My second speculation is that what are deemed NSFW could also be highly aesthetic. If the data is highly curated that every single person looks way above average, prompting “woman” would be the same as prompting “beautiful woman”. Open AI trained the CLIP model with proprietary data. This affects the embeddings of the model. The first suspect is switching from Open AI’s CLIP model to OpenCLIP. Filtered out NSFW contents in training data.This is an area I can only speculate… but why not? The two changes in v2 are Why does negative prompt become more important in v2? It is as if v1.5 model does not understand these words. Adding the negative prompt ugly, deformed and disfigured may improve things but it is not as clear as in v2.1. The images comes out pretty good without any negative prompts in v1.5. Later steps are only finer adjustment to details, such as hairs covering ears.Īdding negative prompt to v1.5. The reasoning behind this is that the diffusion process is most important in the beginning steps. Since, I am using 20 sampling steps, what this means is using the as the negative prompt in steps 1 – 10, and (ear:1.9) in steps 11-20. You will get the same image as if you didn’t put anything. You can verify its uselessness by putting it in the negative prompt. ![]() Let’s pick the as the meaningless, dud negative prompt. You can use keyword switching to first use a meaningless word as negative prompt, and then switching to (ear:1.9) at a later sampling step. Now what if you really want to use a high emphasis (ear:1.9)? I don’t know what’s your problem with ears but I have a trick for you. Negative prompt could affect the diffusion process strongly. The ears are covered more by hair with in all emphasis factors but when the factor reaches 1.9, the composition of the image changed. Below are with three increasing emphasis 1.3, 1.6 and 1.9. What if we are ok with the wind but want the hair to cover the ear? Let’s add negative prompt “ear” with different emphasis factors.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |