Anima Artist Cross-Attn Mix

Blend multiple artist tags for the Anima model so their styles mix evenly instead of one artist dominating the others. Tag artists in the text box with @name or @(name:2.0) (Anima usually wants weights of 2.0+). Each artist is encoded separately and the cross-attention outputs are blended by weight. Built-in NegPiP: any non-artist tag with a negative weight, e.g. (speech bubble:-1.1), is subtracted - no ComfyUI-ppm or extra node needed. Wire both outputs (model + conditioning) into your sampler.