8000 ステップ、19 エポックでトレーニングされたモデル, 顔の品質を高めるには Adetailer が必要です,必要に応じて Lora のウェイトを下げます,
Model trained with 8000 steps and 19 epochs, requires Adtailer to improve facial quality. Lower Lora's weight if necessary.
描述:
The quality of the images has been improved, the training is dim 128, alpha 64,. I can't find the optimal configuration for the alpha. I have not used ADetailer, the result can improve if it is used. But for the first tests I have not used it.
训练词语:
名称: SikiAtsumiV2-10.safetensors
大小 (KB): 147574
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
