# Anima EyesHQ — Hi-Res Detailer V2
I think I can finally call this workflow actually finished? maybe....
A lot of testing went into improving the overall detail quality, especially faces and eyes, while keeping it usable on an RTX 4070 8GB.
⏱️ Around 150–200s per generation on my laptop
At this point, I’m mostly updating it because I apparently don’t know how to stop tweaking things💀
Quality first. Always !
Required Custom Nodes
Impact Pack — https://github.com/comfyorg/comfyui-impact-pack
Impact Subpack — necessary for
UltralyticsDetectorProviderrgthree-comfy — https://github.com/rgthree/rgthree-comfy
JPS Nodes — https://github.com/JPS-GER/ComfyUI_JPS-Nodes
UltimateSDUpscale — https://github.com/ssitu/ComfyUI_UltimateSDUpscale
NVIDIA RTX Nodes — https://github.com/Comfy-Org/Nvidia_RTX_Nodes_ComfyUI
Required Models
Anima Aesthetic 1.0 —
anima_aestheticV10.safetensorsQwen 3 0.6B Base —
qwen_3_06b_base.safetensorsQwen Image VAE —
qwen_image_vae.safetensors4x-UltraSharp —
4x-UltraSharp.pthSAM ViT-B —
sam_vit_b_01ec64.pthAnzhc Face v4 —
Anzhc_Face_v4_y11n.ptEyeful v2 Paired —
Eyeful_v2-Paired.ptNTD11 Anime NSFW Segmentation —
ntd11_anime_nsfw_segm_v5-variant1.pt
UltralyticsDetectorProvider is no longer included directly in Impact Pack, you need to get Impact Subpack separately for recent versions⚠️ RTX Nodes are optional if you don't have a compatible NVIDIA RTX GPU. Simply bypass the RTX upscale group :)
Description
Even more precise but more demanding—tested on an RTX 5060, and the quality is top-notch
I suppose the rendering is just as good on my latest Loras (mainly tested on my Pokemon-style )
