Wan 2.2 (T2V) GGUF + 4 STEP Lightning LORA Workflow

Use this workflow for Wan 2.2 text-to-video: provide text and six named model files, then process the result with the Wan Video 2.2 T2V-A14B base model, a 4-step Lightning LoRA setup, VAE decoding, and video combining.

The 16-node graph covers text encoding, video-latent setup, sampling, VAE decoding, and video combining.

The text path uses CLIPLoader and CLIPTextEncode.

The video path uses EmptyHunyuanLatentVideo, VAEDecode, and VHS_VideoCombine.

Find the workflow on Civitai at https://civitai.com/models/1941674.

The six listed model files are umt5_xxl_fp8_e4m3fn_scaled.safetensors; wan_2.1_vae.safetensors; Wan2.2-Lightning_T2V-v1.1-A14B-4steps-lora_HIGH_fp16.safetensors; Wan2.2-Lightning_T2V-v1.1-A14B-4steps-lora_LOW_fp16.safetensors; Wan2.2-T2V-A14B-HighNoise-Q5_1.gguf; and Wan2.2-T2V-A

The listed node packs are comfyui-gguf and comfyui-videohelpersuite.

The title specifies a GGUF setup with 4 STEP Lightning LORA.

comfyui-videohelpersuite is listed at version 1.6.1.

Before use, verify that all six model files are present under the exact names listed.

Vorschlag · nicht geprüft

Before setup, verify that both listed node packs are available.

Vorschlag · nicht geprüft