WAN 2.2 Speech to Video Comfyui workflow

Szükséges modellek letöltése:

➤ Checkpoints:
wan2.2_s2v_model.safetensors
Place in: /ComfyUI/models/diffusion_models

Wan s2v full model (video.ozeki.hu esetén): hf download Wan-AI/Wan2.2-S2V-14B --local-dir ./

Wan s2v fp8 model (img.ozeki.hu esetén): wget https://huggingface.co/Kijai/WanVideo_comfy_fp8_scaled/resolve/main/S2V/Wan2_2-S2V-14B_fp8_e4m3fn_scaled_KJ.safetensors

mv Wan2_2-S2V-14B_fp8_e4m3fn_scaled_KJ.safetensors wan2.2_s2v_model.safetensors

➤ Audio Encoders:
wav2vec2_large_english_fp16.safetensors
Place in: /ComfyUI/models/audio_encoders

wget https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/resolve/main/split_files/audio_encoders/wav2vec2_large_english_fp16.safetensors

➤ Text Encoders:
native_umt5_xxl_fp8_e4m3fn_scaled.safetensors
Place in: /ComfyUI/models/text_encoders

wget https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/text_encoders/umt5_xxl_fp8_e4m3fn_scaled.safetensors

➤ VAE:
native_wan_2.1_vae.safetensors
Place in: /ComfyUI/models/vae

wget https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/vae/wan_2.1_vae.safetensors

Wan full modell esetén ezzel a kóddal lehet egyesíteni a safetensor fileokat:

nano combine.py

from safetensors.torch import load_file, save_file
import torch
import glob
import os

# Fájlok listája
split_files = sorted(glob.glob("diffusion_pytorch_model-*.safetensors"))

# Összevonás
combined_state_dict = {}
for file in split_files:
    state_dict = load_file(file)
    combined_state_dict.update(state_dict)

# Mentés egyetlen fájlba
output_file = "wan2.2_s2v_model.safetensors"
save_file(combined_state_dict, output_file)
print(f"Összevonva: {output_file}")

Szükséges custom node-ok:
ComfyUI-VideoHelperSuite

Telepítés:

Comfyui Manager menü-> Manage custom nodes -> Search:VideoHelperSuite -> Install

WAN 2.2 Speech to Video Comfyui workflow

More information