MOTION FORGE WAN 2.2 14B T2I (Beginner-Friendly)

Required Models & Exact Locations

Main Checkpoint:

• Wan2.2-T2V-A14B-LowNoise-Q8_0.gguf → ComfyUI/models/unet/

LoRA Files (all optional, can toggle on/off):

• lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16.safetensors → ComfyUI/models/loras/wan_loras/ (Strength: 2.0)

• dslr_video_00003.safetensors → ComfyUI/models/loras/wan_loras/ (Strength: 1.0)

• Vintage-Cinematic-Redmond-WAN22.safetensors → ComfyUI/models/loras/wan_loras/ (Disabled by default)

  • ⚠️ Important: Create subfolder wan_loras inside loras directory

VAE:

• pig_wan_vae_fp32-f16.gguf → ComfyUI/models/vae/

Text Encoder (CLIP):

• cow-umt5xxl-q8_0.gguf → ComfyUI/models/text_encoders/

Upscale Model (optional):

• RealESRGAN_x2plus.pth → ComfyUI/models/upscale_models/

Required Custom Nodes:

1. ComfyUI-GGUF

2. rgthree-comfy (Power Lora Loader, Image Comparer)

3. ComfyUI-Easy-Use

4. ComfyUI_Memory_Cleanup

5. ComfyUI-KJNodes (WAN Video NAG)

6. ControlAltAI-Nodes (FluxResolutionNode)

7. ComfyUI WAN Video Nodes (core)

Workflow Features:

• Purpose: Generate high-quality images/videos from text prompts

• Resolution: Flexible (768x768 default, multiple presets available)

• Steps: 8 (optimized for speed)

• VRAM: 10-14GB recommended

• Beginner Friendly: Power Lora Loader allows easy LoRA toggling

• Includes: 2x upscaling with before/after comparison

Shared Models Between Workflows

These models are used by both workflows (only need one copy):

• pig_wan_vae_fp32-f16.gguf (VAE)

• cow-umt5xxl-q8_0.gguf (CLIP)

• lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16.safetensors (LoRA)

Complete Folder Structure

ComfyUI/

├── models/

│ ├── unet/

│ │ ├── HuMo-17b-Q8_0.gguf

│ │ └── Wan2.2-T2V-A14B-LowNoise-Q8_0.gguf

│ ├── loras/

│ │ └── wan_loras/ ← CREATE THIS SUBFOLDER

│ │ ├── lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16.safetensors

│ │ ├── dslr_video_00003.safetensors

│ │ └── Vintage-Cinematic-Redmond-WAN22.safetensors

│ ├── vae/

│ │ └── pig_wan_vae_fp32-f16.gguf

│ ├── text_encoders/

│ │ └── cow-umt5xxl-q8_0.gguf

│ ├── audio_encoders/

│ │ └── whisper_large_v3_fp16.safetensors

│ └── upscale_models/

│ └── RealESRGAN_x2plus.pth