v2 has been tuned to work well with the standard Euler sampler and low CFG scale. This significantly reduces the color burning that commonly occurs at high CFG values, resulting in more natural color expression.
Compared to v1 on the right, the output from anima_pencil-XL_clear_v2 on the left has more vibrant and lively colors.
On the other hand, to create a cleaner look, the backgrounds have been simplified compared to v1.
Recommended Settings (v2)
CFG scale: 2-4
Sampler: Euler
Scheduler: beta
Steps: 12-20 (around 15)
Quality Tags and Negative Prompts are not always necessary.
This is a lightweight version of v2. The CLIP and VAE remain unchanged, while only the intermediate layers of the UNet have been converted to INT8_ConvRot format.
Image quality is slightly lower than the original v2, but it is particularly well-suited for fast generation in low-VRAM environments (4-6 GB).
INT8_ConvRot is currently supported in ComfyUI and Stable Diffusion webUI Forge NEO.
