This custom node allow controlling output without training. The reducing method is similar to [a/Spatial-Reduction Attention](https://paperswithcode.com/method/spatial-reduction-attention), but generating speed may not be increased on typical image sizes due to overheads. (In some cases, slightly slower)
https://github.com/MitoshiroPJ/comfyui_slothful_attention
git clone https://github.com/MitoshiroPJ/comfyui_slothful_attention
comfy node install comfyui_slothful_attention