Minimalist T2I Workflow for WAN Video 2.2

Minimalist T2I Workflow for WAN Video 2.2 is for generating static images from text prompts with WAN Video 2.2 high- and low-noise 14B FP8 models.

The listed components include PreviewImage and SaveImage for the described static-image output.

Input: a text prompt describing the image to generate.

Output: a static image, with PreviewImage and SaveImage listed among the components.

GegenDenTag publishes it on Civitai at https://civitai.com/models/2163207.

The page classifies it as Workflows and lists Wan Video 2.2 T2V-A14B as the base model.

The listed model files are umt5_xxl_fp8_e4m3fn_scaled.safetensors, wan_2.1_vae.safetensors, wan2.2_t2v_high_noise_14B_fp8_scaled.safetensors, and wan2.2_t2v_low_noise_14B_fp8_scaled.safetensors.

Core component names include CLIPLoader, CLIPTextEncode, EmptySD3LatentImage, GetNode, KSampler, ModelSamplingSD3, SetNode, and UNETLoader.

Additional component names are MarkdownNote, KSamplerAdvanced, PreviewImage, SaveImage, VAEDecode, and VAELoader.

The published setup omits Lighting LoRAs and additional conditioning through LoRAs and CNet.

It is described for evaluating how the base models represent artistic styles, techniques, and compositions.

Before using it, verify that your environment includes the named model files and components, including wan_2.1_vae.safetensors.

Vorschlag · nicht geprüft

No release version is provided for this workflow.