我使用的加速LORA并且步骤=12与CFG=1，当然这不是绝对的，我没有过多的尝试其他的情况。
I am using acceleration LORA with steps=12 and CFG=1, but this is not absolute as I have not tried many other scenarios.
If you are using the anima-turbo-lora-v0.1 acceleration LoRA, please use CFG 1 and Step 12. These settings are specifically designed to work with this LoRA's optimization.
If you are NOT using this acceleration LoRA and only using the ANIMA BASE 1.0 main model, I have not tested other parameters. Please feel free to experiment and find the settings that work best for you.
Description
我使用的加速LORA并且步骤=12与CFG=1，当然这不是绝对的，我没有过多的尝试其他的情况。
I am using acceleration LORA with steps=12 and CFG=1, but this is not absolute as I have not tried many other scenarios.
If you are using the anima-turbo-lora-v0.1 acceleration LoRA, please use CFG 1 and Step 12. These settings are specifically designed to work with this LoRA's optimization.
If you are NOT using this acceleration LoRA and only using the ANIMA BASE 1.0 main model, I have not tested other parameters. Please feel free to experiment and find the settings that work best for you.
Comments (4)
Please provide a full description of the image to get the same result.
ANIM BASE1.0+anima-turbo-lora+Step 12 CFG 1 My prompt is simple, a woman with ancient Chinese style
@TwoCat_MUGEN_HELL I say this because you're posting lore, not a base model or just a generated image. You need to know and rely on some kind of result and what you used to achieve your result. In most cases, modified workflows are used instead of the stock ones, and this is simply terrible.
Thank You.
@alexxz Thanks for your feedback. I think you might be missing the key point here: CFG 1 and Step 12 are only possible because of the acceleration LoRA (anima-turbo-lora-v0.1). Without this LoRA, ANIM BASE 1.0 would require much higher CFG and steps. This is a specific optimized workflow, not a 'terrible' default setup. It works exactly as intended.