Donate Coffee for Gtonero

Not recommended for realistic model

Since training with Conv2D requires
https://github.com/KohakuBlueleaf/a1111-sd-webui-locon

Tag : blush, embarrassed,

Training Data v.1.0

• 125 image 10 repeats 3 epochs

• Powershell ISE run script kohya-ss

• module = Networks.LoRA

• cosine_with_restarts 4

• 10 DIM / 5 Alpha | ConvDIM/Alpha 4

• U-net 1e-4 / Encoder 2.5e-5

• Loss 0.0767