After merging a checkpoint, I often train a LoRA using a familiar character to test how the checkpoint performs and to compare it with others. Recently, I’ve been focusing on Jin Yuichi, which is why there are multiple LoRAs of him — he’s my current benchmark character (01.2026).

Jin Yuichi (World Trigger) - My current benchmark character (01.2026).

Change training model to WAIMIX-1.5

jin-yuichi-waimix-15.safetensors

WAIMIX-1.5

BASIC_ILL

Files with SHA256: ac681d2d8fe79d2da6f0c0c87749eb74894d5c05ffe1abdbe7eb0a284b04d1d9