Breakthrough efficiency for Sound-to-Video. This revolutionary workflow runs the massive 14B parameter WAN2.2 S2V model on consumer hardware by leveraging fully quantized GGUF models for the UNET and CLIP. Experience true audio-driven animation with drastically reduced VRAM requirements, making high-end S2V generation accessible to everyone. CPU/GPU hybrid execution enabled.<hr /><h2 id="workflow-description">Workflow Description</h2>This workflow is a technical marvel, designed to democratize access to the powerful WAN2.2 Sound-to-Video 14B model. By utilizing the ComfyUI-GGUF plugin, it loads both the UNET and CLIP models in highly compressed, quantized GGUF formats. This translates to:<ul><li>Massive VRAM Savings: The Q2_K quantized UNET allows the 14B model to run on GPUs with as little as 8-10GB of VRAM, or even on capable CPU systems.</li><li>Hybrid Execution: Seamlessly offloads layers between GPU and CPU to maximize performance on any hardware setup.</li><li>Full Fidelity Functionality: Despite the compression, you retain the complete S2V feature set: audio-driven motion, high-quality output, and professional video encoding.</li></ul>This is the ultimate solution for users who thought the 14B S2V model was out of reach. Now you can run it.<hr /><h2 id="features-and-technical-details">Features &amp; Technical Details</h2>🧩 The Quantized Stack (The Magic Sauce):<ul><li>UNET (GGUF): <code>Wan2.2-S2V-14B-Q2_K.gguf</code> - The core video generation model, quantized to 2-bit for extreme efficiency.</li><li>CLIP (GGUF): <code>umt5-xxl-encoder-q4_k_m.gguf</code> - The text encoder, quantized to 4-bit for optimal performance.</li><li>VAE: <code>Wan2.1_VAE.safetensors</code> - Loaded normally for maximum visual fidelity.</li><li>Audio Encoder: <code>wav2vec2_large_english.safetensors</code> - Encodes the input audio for the model.</li></ul>🎬 Core Functionality:<ul><li>True Sound-to-Video: The generated animation is directly influenced by the characteristics of your input audio.</li><li>Auto-Duration: Automatically calculates the correct number of video frames (<code>length</code>) based on your audio file's duration.</li><li>Smart Image Preprocessing: Automatically scales your input image to an optimal size (0.2 megapixels) while preserving its original aspect ratio for animation.</li><li>Professional Output: Uses <code>VHS_VideoCombine</code> to render a final MP4 video with perfect audio sync.</li></ul>⚙️ Optimized Pipeline:<ul><li>Clean, grouped node layout for easy understanding and operation.</li><li>Efficient routing with reroute nodes to keep the workflow organized.</li></ul><hr /><h2 id="how-to-use-steps-to-run">How to Use / Steps to Run</h2><h3 id="prerequisites-(critical):">Prerequisites (CRITICAL):</h3><ol><li>ComfyUI-GGUF Plugin: You MUST install the <code>ComfyUI-GGUF</code> plugin from its GitHub repository. This is non-negotiable.</li><li>GGUF Model Files: Download the required quantized models:<ul><li><code>Wan2.2-S2V-14B-Q2_K.gguf</code> (Place in <code>Qwen 2SV\</code> folder)</li><li><code>umt5-xxl-encoder-q4_k_m.gguf</code></li></ul></li><li>Standard Models: Ensure you have <code>Wan2.1_VAE.safetensors</code> and <code>wav2vec2_large_english.safetensors</code>.</li></ol><h3 id="instructions:">Instructions:</h3><ol><li>Load Your Image: In the <code>LoadImage</code> node, select your starting image.</li><li>Load Your Audio: In the <code>LoadAudio</code> node, select your <code>.wav</code> or <code>.mp3</code> file.</li><li>Craft Your Prompt: Describe your scene in the Positive Prompt node. The negative prompt is pre-configured.</li><li>Queue Prompt. The workflow will encode your audio, process it through the quantized 14B model, and generate the video.</li></ol>⏯️ Output: Your finished video will be saved in your ComfyUI <code>output/video/</code> folder as an MP4 file.

Files with SHA256: 87e9aa4246517fc8b6d4d0c84cb0623c5b3dd878f7615d68225577b699f889db