Type | |
Stats | 443 279 326 |
Reviews | (124) |
Published | Jun 2, 2025 |
Base Model | |
Training | Steps: 5,550 Epochs: 80 |
Usage Tips | Clip Skip: 2 |
Trigger Words | pixel-art pixel pixel_art |
Hash | AutoV2 AE89C5F322 |
The training dataset consisted of artwork where each pixel was equal to 8x8. This is important because if you need "perfect" pixels, I recommend RESIZE at 0.125, and then RESIZE at 8.000, so that the pixels acquire their true nature. Use the Nearest Neighbor method when resizing.
We are all waiting for the new versions together. According to the plans: to get rid of harsh inaccuracies and broken pixels (I will probably achieve this with high DIM and ALPHA). I would like to keep dithering under control, but almost all the artworks in the training include dithering, so I can't say what will come of it.
Changelog
Version 2
I expanded the dataset to 281 images (there were about 50 in the first version). The variety of possible generations has increased significantly, but the pixels are still not perfect. And you know what? I intend to review absolutely all the images in the dataset because, in my opinion, the problem might lie in hidden imperfections. If even one of the images has a slight gradient or a broken pixel, the entire training process could be ruined. But what terrifies me even more is that I can't find a similar "pixelate+" tool. For example, in paint.net, there is a "pixelate+" effect, and the pixelation method there is so good that the image is perfectly pixelated almost without changing. And I'm very sorry that I haven't been able to find a similar tool in Comfy yet. If you use the RESIZE method like I do, I have bad news for you—this method is not ideal. At least, that's how it was in my tests. Our conditions might differ, and in reality, everything might work out great for you.
Version 1 (test)
Frankly, it turned out not quite what I want to achieve in the next version.
You can wait for the release of version 2, it will be more worthy than the current one.
If you are not scared by the words above, then know that the test LORA is strange, over-trained and tends to draw much more often what it was trained on than to match text prompts, keep this in mind.
You will encounter many swords, strange books, Easter Island statues during exploitation and, most importantly: GigaChad (even a girl can acquire his facial features, but this is uncontrollable (in any case, I didn't try)).