V2:PLORA，target_res1024。

针对此案例1536，在非TURBO出图时会有笔触过深的问题，所以换回1024，同时修改了配方，在非使用TURBO LORA时不会笔触过重。但相应地风格没有那么强烈，如果你配搭TURBO lora用，请下载V1。

V2: P LoRA, target_res 1024.

1536 refers to the training resolution. V1 was trained at 1536. For this V2 upgrade, training at 1536 caused issues in this case, so I went back to 1024 for training and also adjusted the recipe. As a result, outputs no longer have overly heavy brushstrokes when the TURBO LoRA is not used. The trade-off is that the style is less pronounced. If you plan to pair it with the TURBO LoRA, please download V1.

基于base1.0训练，T-LORA训练，用target_res 1536训练。

触发词:@knastyle

LORA权重:0.8~1

Sampler: er_sde + beta

演示图依然搭配了Anima Turbo LoRA v0.2 ，但V2没有。

Trained based on base1.0, T-LORA training, trained with target_res 1536.

Trigger word: @knastyle

LoRA weight: 0.8~1

Sampler: er_sde + beta

showcases are paired with Anima Turbo LoRA v0.2,without V2.