Prompt
2d icon. {your prompt}. <lora:game_icon_v1.0:1>The shorter the prompt, the better. Use the "2d icon" modifier at the beginning. Then briefly describe the object. Put Lora at the end of the prompt.
Example:
2d icon. a basket of lemons with leaves on top. <lora:game_icon_v1.0:1>
2d icon. a blue skirt with ruffles. <lora:game_icon_v1.0:1>
2d icon. a bottle of oil with a cork stopper. <lora:game_icon_v1.0:1>
2d icon. a mushroom with a white base and orange cap. <lora:game_icon_v1.0:1>
Negative prompt
I recommend using this one: «(blurry:1.3). lowres.».
___________________
Промпт
2d icon. {ваш промпт}. <lora:game_icon_v1.0:1> Чем короче промпт – тем лучше. В начале используйте модификатор «2d icon». Затем коротко опишите объект. Lora ставьте в конец промпта.
Примеры:
2d icon. a basket of lemons with leaves on top. <lora:game_icon_v1.0:1>
2d icon. a blue skirt with ruffles. <lora:game_icon_v1.0:1>
2d icon. a bottle of oil with a cork stopper. <lora:game_icon_v1.0:1>
2d icon. a mushroom with a white base and orange cap. <lora:game_icon_v1.0:1>
Негативный промпт
Рекомендую использовать этот: «(blurry:1.3). lowres.».
___________________
www.t.me/alchemist_tm – Мой телеграм-канал
https://youtube.com/@TechnoMagix – Мой ютюб
https://t.me/+EOhMG42mlywxNWMy – Чат по нейросетям. Взаимопомощь и творчество.
https://sponsr.ru/technomagix - открытый курс по SD
Description
FAQ
Comments (20)
Cool! Thanks for sharing, thats amazing stuff)
you're welcome
Amazing! Could your please share your training details?
I plan to record a video tutorial where I will share my settings and show many tables with results for different training settings.
@technomagix Thank you! Waitting for your video!
К сожалению что-то не работает.
Вначале никак не хотелось скачиваться. Получилось с попытки 10ой, но даже после этого, при переносе файла в папку Lora, она так и не появляется в программе :(
Лора совместима только с SDXL. Чтобы лора была видна через программу нужно сначала в программе сделать активным чекпоинт SDXL. Если будет загружен чекпоинт стейбла версии 1 или 2, то программа автоматически скроет неподходящие лоры.
По поводу скачивания ничего сказать не могу, это от сайта зависит, я над ним не властен.
p.s. Удобнее и быстрее получить ответ в телеграм-чате https://t.me/+EOhMG42mlywxNWMy
Sorry for mention again, did you share instruction video for training this lora somewhere, I really like this model and want to know more details!
新手迷惑 这是Lora？这么大，印象里lora都是M
Yes, this is lora. I have a version that weighs 2 times less, but it doesn’t give worse results, so I didn’t post it.
It seems not to working with SDXL in EasyDiffusion?
"Error: Could not load the lora model! Reason: Sorry, could not load D:\EasyDiffusion\models\lora\game_icon_XL_v1.0.safetensors. You're trying to use a SD1 LoRA model with a SDXL Stable Diffusion model. They're not compatible, please use a compatible model!"
is this model available on diffusers?
I failed to apply it to diffusers
为什么感觉还原不出这么好的效果，真的只是简单描述就可以吗？
Hi, I plan to train a similar icon LoRA based on SDXL using my own dataset, but the resolution for the data is far below 1024. Have you faced such a problem? Is there any recommendation for it except upscaling the pictures? Thank you in advance.
i can't get it to work as a lora is it normal ? my a1111 doesn't detect it
Pretty good, can you make lora for web & apps icons?
I am sad as I cannot replicate any of these excellent images using this LORA. All my images turn out like a child's drawing :-(
Any guide on how to get the Lora to work in automatic1111? I haven't managed to do it, so far...
in models there will be a folder called lora, just copy paste the model in there, then when you open automatic 1111 you can see a lora tab, all the lora you copied you can see on there, to use lora you need a check point, some Loar required specific model, make sure you have a checkpoint
搭配底模pixelwaveturboExcellent_03可以在SD-forge版本生成透明通道图片，<lora:SDXL_icon_V2:1>的效果就不错，感谢