
Today I woke up with an exhausting cold, I had a problem with training a lora and the brightness on the screen kept me away from the PC (I always use less than half so it was a problem).
I guess filling up on vitamin C had an effect, although this would have been taken away with some cold pills but I was too lazy to get out of bed. XD
So... now that I'm better I could work on that lora already mentioned before but I'm too lazy to organize so many images. And while reviewing the bounties I saw someone asking for an image of a similar style and I thought: maybe you can't create loras with an image ?
To avoid overfitting you just need to break the images, rotate them and crop them. (Although in this case it is a problem because the reference file was an image of Garfield, but lowering it to 0.7 works fine, what interests me are the brush strokes).
Anyway, it was interesting to create a style with an image :3
描述:
训练词语: totest_st
名称: 513374_training_data.zip
大小 (KB): 3429
类型: Training Data
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
名称: HerrscherAGGA2024_StyleToTest_V1.safetensors
大小 (KB): 36980
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success