WAN 2.2 T2I

Description:
This workflow utilizes the latest Wan 2.2 (14B) models in GGUF format to generate high-quality vertical videos (1088x1440) from text prompts. It features a sophisticated dual-stage sampling process:

1. High Noise Pass: Uses the Wan2.2-T2V-A14B-HighNoise model to establish the base composition and movement.
2. Low Noise Pass: Switches to the Wan2.2-T2V-A14B-LowNoise model to refine details and textures.

This GGUF implementation allows running the massive 14B model with reduced VRAM requirements while maintaining high fidelity. The setup also includes a chain of LoRAs for specific styles (Instagirl, Lenovo) and performance optimization (Distill).

Key Features:

• GGUF Optimization: Runs 14B models efficiently using Q8_0 quantization.
• Dual-Model Sampling: Splits generation between High-Noise and Low-Noise checkpoints for superior quality.
• LoRA Stack: Pre-configured with a "Distill" LoRA for faster steps and custom style LoRAs (Instagirl, Lenovo).
• Vertical Ratio: Default set to 1088x1440, perfect for social media content.

Links to my patreon and youtube:

https://www.patreon.com/cw/ASF_community?utm_medium=unknown&utm_source=join_link&utm_campaign=creatorshare_creator&utm_content=copyLink

https://www.youtube.com/@ASF_One-Click-PRO