wan/2.2/face-detailer/README.md

WAN 2.2 Face Detailer

Model: WAN 2.2 T2V A14B (fp8 scaled) Type: Video-to-Video

Face enhancement pipeline for WAN 2.2 video output. Loads an existing video, detects and segments faces using SAM2, crops each face region, runs a focused WAN 2.2 generation pass to add detail, then composites the result back into the original frame.

---

Preview

---

Requirements

• ComfyUI: recent stable build

Models

• Main model: Wan2_2-T2V-A14B-LOW_fp8_e4m3fn_scaled_KJ.safetensors — Kijai/WanVideo_comfy_fp8_scaled

• VACE module: Wan2_2_Fun_VACE_module_A14B_LOW_fp8_e4m3fn_scaled_KJ.safetensors — Kijai/WanVideo_comfy_fp8_scaled

• VAE: Wan2_1_VAE_bf16.safetensors — Kijai/WanVideo_comfy

• Text encoder: umt5-xxl-enc-bf16.safetensors — Kijai/WanVideo_comfy

• LoRA (step distill): lightx2v_T2V_14B_cfg_step_distill_v2_lora_rank64_bf16.safetensors — Kijai/WanVideo_comfy

• LoRA (reward): Wan2.2-Fun-A14B-InP-low-noise-HPS2.1.safetensors — example LoRA, replace with your own if needed — alibaba-pai/Wan2.2-Fun-Reward-LoRAs

• Segmentation: sam2.1_hiera_small.safetensors — Kijai/sam2-safetensors

Custom nodes

• ComfyUI-WanVideoWrapper
• ComfyUI-KJNodes
• ComfyUI-VideoHelperSuite
• ComfyUI-segment-anything-2
• rgthree-comfy
• ComfyUI-Florence2 — delete these nodes if you don't want to use florence2 for segmentation.

---

Notes

• The VACE module is optional — disconnect it to run without VACE.
• SAM2 is used for face segmentation. The face bounding box index can be adjusted directly in the workflow if the wrong face is selected.

---

Changelog

• 2026-03-29 — Initial upload