Wan 2.2 S2V in ComfyUI: Talking Images (GGUF Model) Workflow — v1.0

Use this ComfyUI workflow to make a talking-image video from an image and audio; it includes image-path and audio input nodes and a video-combine node.

Its listed nodes cover audio and text encoding, Wan sound-image-to-video processing, image resizing, sampling, VAE decoding, and video assembly.

Use VHS_LoadImagePath for the image path and VHS_LoadAudio for audio; VHS_VideoCombine is the named video-assembly node.

The listing is version v1.0 by Lordcaocao2025, with base model recorded as Other.

Find the workflow on Civitai: https://civitai.com/models/2514688?modelVersionId=2826518

The workflow lists these model files: - lightx2v_T2V_14B_cfg_step_distill_v2_lora_rank64_bf16.safetensors - umt5_xxl_fp8_e4m3fn_scaled.safetensors - wan_2.1_vae.safetensors - Wan2.2-S2V-14B-Q4_K_S.gguf - wav2vec2_large_english_fp16.safetensors

The listed node packs are: - comfyui_essentials - comfyui-art-venture - comfyui-custom-scripts - comfyui-gguf (version 1.1.10) - comfyui-videohelpersuite - ComfyUI-WanVideoWrapper

For prompt, sampling, LoRA, or batch settings, inspect CLIPTextEncode, KSampler, LoraLoaderModelOnly, ModelSamplingSD3, and BatchCount+.

Vorschlag · nicht geprüft

For image sizing and video assembly, inspect WanVideoImageResizeToClosest and VHS_VideoCombine.

Vorschlag · nicht geprüft

Locate Wan2.2-S2V-14B-Q4_K_S.gguf before setup; the listing names the file but gives no download URL.

Vorschlag · nicht geprüft

Confirm which installed package provides comfyui-gguf; the workflow lists that name at version 1.1.10.

Vorschlag · nicht geprüft