Install Wan Video

This workflow is for installing Wan Video and supports image-to-video generation: it takes an image and text and lists a WEBM video as its output.

The listed 17-node flow loads an image, encodes vision and text inputs, runs Wan image-to-video processing and sampling, decodes with a tiled VAE, and saves WEBM.

The meaningful inputs are an image and text; the listed output format is WEBM video.

The workflow is published by Tingtingin and uses Wan Video as its base model.

Required model files are `clip_vision_h.safetensors`, `umt5_xxl_fp8_e4m3fn_scaled.safetensors`, `wan_2.1_vae.safetensors`, and `wan2.1-i2v-14b-480p-Q4_K_S.gguf`.

The setup also lists a node pack named `comfyui-gguf`.

The listed model details use Wan 2.1 I2V 14B with GGUF, FP8, Q4, and Q4_K labels.

Before use, check that all four model files and `comfyui-gguf` are available under the listed names.

Sugerencia · sin verificar

No license is specified for this workflow.