{ "id": 19075, "modelId": 16014, "name": "v2.0 offset", "createdAt": "2023-03-05T20:00:31.679Z", "updatedAt": "2023-03-25T13:40:00.364Z", "trainedWords": [ "monochrome", "lineart" ], "baseModel": "SD 1.5", "earlyAccessTimeFrame": 0, "description": "
V2.0 is updated! Trained with offset noise, and all demos are generated with Anything V4.5 base model and orange mix VAE.
What this model does: to generate true linearts! (not that manga-like monochrome illustrations)
You can use it without trigger words, but sometimes there will be some unwanted colors, you can use monochrome
and lineart
to avoid that! (also, remove all colors from your prompt if you don't bother)
And, with the help of ControlNet, you can convert any image into lineart. I made a video to describe the process. If the line is too thin, try lowering the low threshold for the canny preprocessor. Also, you can use multi-ControlNet to assist.
Try applying prompts like sketch
or doodles
, and you can get a cleaned sketch; by applying chartuner LoRA and openpose helper image, you can get a decent character reference sheet ... this LoRA is full of possibilities!
But sometimes it doesn't go well with character LoRAs, you can either lower the strength of the character LoRA or this LoRA. This LoRA should work fine with a strength of 0.8~1.
Note: since this LoRA is trained based on Anything V4.5, using other base models may have some deteriorations, and you may need to adjust the strength for different base models. Check this reference sheet for different models:
V2.0\u5df2\u66f4\u65b0\uff0c\u57fa\u4e8eoffset noise\u8bad\u7ec3\uff0c\u6240\u6709\u5c55\u793a\u56fe\u7247\u5747\u4f7f\u7528Anything V4.5\u4f5c\u4e3a\u57fa\u6a21\uff0c\u5e76\u4f7f\u7528orange mix\u7684VAE\u3002
\u672c\u6a21\u578b\u76ee\u7684\uff1a\u751f\u6210\u5e72\u51c0\u7684\u7ebf\u7a3f\u56fe\uff0c\u800c\u975e\u6f2b\u753b\u98ce\u9ed1\u677f\u56fe\u3002
\u6b63\u5e38\u65f6\u65e0\u9700\u4f7f\u7528\u89e6\u53d1\u8bcd\uff0c\u4f46\u6709\u65f6\u4f1a\u51fa\u73b0\u4e0d\u60f3\u8981\u7684\u989c\u8272\uff0c\u53ef\u4ee5\u4f7f\u7528 monochrome
\u6216 lineart
\u6765\u52a0\u4ee5\u8f85\u52a9\u3002
\u8f85\u4ee5ControlNet\uff0c\u5373\u53ef\u5c06\u4efb\u4f55\u56fe\u7247\u5747\u8f6c\u5316\u4e3a\u7ebf\u7a3f\u98ce\u683c\uff0c\u6211\u505a\u4e86\u4e00\u4e2a\u89c6\u9891\u6765\u7b80\u5355\u8bb2\u89e3\u3002\u5982\u679c\u751f\u6210\u7684\u7ebf\u592a\u7ec6\uff0c\u53ef\u4ee5\u5c1d\u8bd5\u964d\u4f4ecanny\u9884\u5904\u7406\u5668\u7684low threshold\uff0c\u4e5f\u53ef\u4f7f\u7528multi-ControlNet\u6765\u8f85\u52a9\u3002
\u4f7f\u7528\u8bf8\u5982 sketch
\u6216 doodles
\u4e4b\u7c7b\u7684prompt\u53ef\u4ee5\u751f\u6210\u66f4\u5e72\u51c0\u7684\u8349\u7a3f\u98ce\u683c\u56fe\u7247\uff1b\u4f7f\u7528chartuner LoRA\u548copenpose helper image\u5219\u53ef\u4ee5\u751f\u6210\u4eba\u7269\u4e09\u89c6\u56fe\u3002\u672cLoRA\u53ef\u73a9\u5ea6\u5f88\u9ad8\uff01
\u6709\u65f6\u4f7f\u7528\u4eba\u7269LoRA\u53ef\u80fd\u4f1a\u51fa\u73b0\u95ee\u9898\uff0c\u53ef\u4ee5\u901a\u8fc7\u964d\u4f4e\u4eba\u7269LoRA\u7684\u6743\u91cd\u6216\u8005\u964d\u4f4e\u672cLoRA\u6743\u91cd\u6765\u8c03\u6574\u6548\u679c\u3002\u672cLoRA\u5f3a\u5ea6\u63a8\u83500.8~1\u3002
\u6ce8\uff1a\u56e0\u672c\u6a21\u578b\u662f\u57fa\u4e8eAnything V4.5\u8fdb\u884c\u8bad\u7ec3\u7684\uff0c\u6545\u4f7f\u7528\u5176\u4ed6\u57fa\u6a21\u65f6\u53ef\u80fd\u6027\u80fd\u4f1a\u6709\u6240\u4e0b\u964d\uff0c\u53ef\u80fd\u9700\u8981\u5bf9\u5f3a\u5ea6\u8fdb\u884c\u8c03\u6574\u3002\u4e0b\u56fe\u4e3a\u4e0d\u540c\u6a21\u578b\u7684\u5f3a\u5ea6\u53c2\u8003\uff1a
https://imagecache.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/6a8a7f93-32ff-4c9d-9689-197ca1ae7600/width=3968/203078