wan-t2v-video

1# WAN 2.2 Text-to-Video (T2V) Workflows
2 
3## Overview
4 
5WAN 2.2 T2V generates videos from text prompts using a 14B parameter MoE (Mixture of Experts) architecture split across two specialized models:
6 
7- HighNoise model handles early denoising. It establishes structure, motion, composition
8- LowNoise model handles late denoising. It refines details, sharpens output
9 
10This dual-model technique is the same as FLF/I2V (see wan-flf-video skill) but without image conditioning nodes.
11 
12Key difference from I2V/FLF: T2V does NOT use `CLIPVisionEncode`, `WanFirstLastFrameToVideo`, or any image input. It uses `EmptyHunyuanLatentVideo` for latent initialization and text-only conditioning.
13 
14## Models
15 
16### UNET (Installed)
17 
18| Model | Loader | Notes |
19|-------|--------|-------|
20| `Wan2_2-T2V-A14B_HIGH_fp8_e4m3fn_scaled_KJ.safetensors` | `UNETLoader` | HighNoise expert, 14.3GB FP8 |
21| `Wan2_2-T2V-A14B-LOW_fp8_e4m3fn_scaled_KJ.safetensors` | `UNETLoader` | LowNoise expert, 14.3GB FP8 |
22 
23### Text Encoder
24 
25| Component | Node | Model | Notes |
26|-----------|------|-------|-------|
27| **CLIP (T5)** | `CLIPLoader` (type=`wan`) | `umt5_xxl_fp8_e4m3fn_scaled.safetensors` | UMT5-XXL fp8, in clip/ |
28 
29### VAE
30 
31| Component | Node | Model |
32|-----------|------|-------|
33| **VAE** | `VAELoader` | `wan_2.1_vae.safetensors` |
34 
35### VACE Modules (Installed — For Advanced Control)
36 
37| Model | Size | Notes |
38|-------|------|-------|
39…