This contains multiple pony v6 models. The weights are adjusted for brighter colors.

Qolor

qolor_v2.safetensors

creator: qeltModels mainly for anime. Each of the versions have a style, but they are compatible with style loras.Score tags or negative prompts are not required.Knows more tags from danbooru such as; artists, characters, styles, concepts.The additional knowledge comes from <a target="_blank" rel="ugc" href="https://civitai.com/models/312010/4th-tail-lora-extract">this lora</a>.Example images are raw output from one ksampler.Nodes for further quality:<a target="_blank" rel="ugc" href="https://github.com/Haoming02/comfyui-diffusion-cg">Latent color grading</a> (pass right after empty latent)<a target="_blank" rel="ugc" href="https://github.com/pamparamm/sd-perturbed-attention">Perturbed attention guidance</a> (set scale to near 1, unet_block to 2 or 3 and output)<a target="_blank" rel="ugc" href="https://github.com/pamparamm/ComfyUI-vectorscope-cc">Vectorscope</a> (useful for dark images)<a target="_blank" rel="ugc" href="https://github.com/mcmonkeyprojects/sd-dynamic-thresholding">Dynamic thresholding</a> (prevent cfg burn) or <a target="_blank" rel="ugc" href="https://github.com/Extraltodeus/ComfyUI-AutomaticCFG">Automatic CFG</a> (no burn)My favorite upscaler is <a target="_blank" rel="ugc" href="https://openmodeldb.info/models/2x-NomosUni-span-multijpg">2x-NomosUni-span-multijpg</a>VAE used in cc version is regular <a target="_blank" rel="ugc" href="https://civitai.com/models/296576/sdxl-vae">sdxl_vae</a>.a &amp; b have <a target="_blank" rel="ugc" href="https://civitai.com/models/152040?modelVersionId=812935">this one</a> included.

Files with SHA256: 854df1999fd30bd26beb8e30ef08d0c68071321832b9f3d5fa365c4c0314bd48