My personal basic and simple Wan2.1/Wan2.2 I2V workflows based on ComfyUI native one

This ComfyUI-native Wan2.1/Wan2.2 I2V workflow is for image-to-video work: it takes image and text inputs, applies image scaling and text encoding, and can create, preview, and save video outputs.

The workflow includes image loading and scaling, text encoding, sampling, decoding, WanImageToVideo processing, video creation, previewing, and saving.

Use LoadImage for image input, CLIPTextEncode for text input, and ImageScale for image scaling.

CreateVideo, PreviewImage, and SaveVideo cover video creation, preview, and saving.

Published by DigitalZombieAI on Civitai at https://civitai.com/models/1389968.

The listing classifies it as Workflows and names Wan Video 2.2 I2V-A14B as the base model.

Named model files include lightx2v_T2V_14B_cfg_step_distill_v2_lora_rank64_bf16.safetensors, umt5_xxl_fp16.safetensors, and wan_2.1_vae.safetensors.

It also names wan2.2_i2v_high_noise_14B_fp8_scaled.safetensors and wan2.2_i2v_low_noise_14B_fp16.safetensors.

Listed node types are CLIPLoader, CLIPTextEncode, CreateVideo, ImageScale, KSamplerAdvanced, LoadImage, LoraLoaderModelOnly, ModelSamplingSD3, PathchSageAttentionKJ, PreviewImage, SaveVideo, UNETLoader, VAEDecode, VAELoader, VRAM_Debug, and WanImageToVideo.

The listing does not specify a license.

Before use, verify that all five named model files are present and have the listed filenames.

Suggestion · non vérifié

Before use, verify that the listed node types are available, including CLIPLoader, SaveVideo, VRAM_Debug, KSamplerAdvanced, and PathchSageAttentionKJ.

Suggestion · non vérifié