<a target="_blank" rel="ugc" href="https://old.reddit.com/r/StableDiffusion/comments/1h7hunp/how_to_run_hunyuanvideo_on_a_single_24gb_vram_card/">Initial comfy setup</a>For v1.01) <a target="_blank" rel="ugc" href="https://huggingface.co/Kijai/WanVideo_comfy/tree/main">Weights to generate</a>2) <a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-WanVideoWrapper">Github the main node</a>For v2.01) <a target="_blank" rel="ugc" href="https://huggingface.co/city96/Wan2.1-I2V-14B-480P-gguf/blob/main/wan2.1-i2v-14b-480p-Q4_K_M.gguf">I2V Quant Model -&gt; </a>models/diffusion_models2) <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/blob/main/split_files/clip_vision/clip_vision_h.safetensors">CLIP VISION -&gt; </a>models/clip_vision3) <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/blob/main/split_files/vae/wan_2.1_vae.safetensors">VAE -&gt; </a>models/vaePrompt can write Florence2 LLM (workflow has toggle between Florence2 and manual user prompt).Tested on 4080 16GB<hr />v2.0 dev in development. If you have problems with AnyPython node, you should switch the channel recent or/and lower security_level (normal- OR weak) in the ComfyUI/Custom_nodes/ComfyUI-Manager/config.json

Files with SHA256: 3f85989580dde4273773b68988f7bd3db107b419e5704915693e9c6528e3d09c