The power of goon can't never be stopped by hardware limitations! In an attempt to force Wan to run on my pc having less than 16GB of VRAM and extremely slow AI GPU I present with you with this caveman easy workflow.<hr />I M P O R T A N T: I have no idea what I'm doing, if you know how to speed this workflow up and how to improve it, let me know right away! I'm using a fresh portable installation.<hr />Requirements:I'm not sure about the requirements, but my setup is using max of 8.1GB of VRAM and 16.2GB of RAM - After the first load, which requires nearly 28GB of RAM.<ul><li>Minimum of 12GB of VRAM - There's no GPU with 10GB or 9GB anyways.</li><li>Minimum of 28GB of RAM - It will not work with 16GB or less, trust me.</li></ul>Features:<ul><li>Basic LoRa Support for now.</li></ul>Models Needed:<ul><li>Model is Rapid Wan AIO 14B SFW - Q4_K_M / <a target="_blank" rel="ugc" href="https://huggingface.co/befox/WAN2.2-14B-Rapid-AllInOne-GGUF/tree/main/Mega-v10">Download Link</a></li><li>Model is Rapid Wan AIO 14B NSFW - Q4_K_M / <a target="_blank" rel="ugc" href="https://huggingface.co/DoorZekor/WAN2.2-14B-Rapid-AllInOne-GGUF-NSFW-v10">Download Link</a></li><li>Vae is Wan 2.1 Vae - Default / <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">Download Link</a></li><li>Text Encoder Clip is umtt5-xxl-encoder - Q5_K_M / <a target="_blank" rel="ugc" href="https://huggingface.co/city96/umt5-xxl-encoder-gguf">Download Link</a></li></ul>Read Me:<ul><li>It is better to run 6 steps rather than 4 for more details, or use a detailer lora.</li><li>Negative chinese prompt is obligatory, otherwise results will be shit, add your own english tags after.</li><li>I use "No camera movement" instead of long sentences.</li><li>"Detailed movement" and other quality tags seems to work.</li><li>It is said that NSFW version is worse than SFW at physics and accurate animation despite understanding more concepts. But SFW looks much worse for me.</li><li>Seed "42" gives better results for static light novel-like animations.</li><li>You can simply bypass Save Video and LoRa loading if needed.</li></ul><hr />Performance:<ul><li>Average of 3-6 minutes for 4 seconds at 800x400 resolution.</li><li>Average of 3-6 minutes for 5 seconds at 700x500 resolution.</li><li>Average of 3 minutes for 5 seconds at 400x400 resolution.</li><li>Average of 56 seconds for 5 seconds at 200x250 resolution.</li></ul><ul><li>Average of ~2 minutes at 416x512 resolution at 6 steps, fastest.</li></ul>Performance Tips:<ul><li>If you can set up Sage Attention you will have extremely faster performance.</li><li>SSD/NVME will be better for RAM file management. Mine is running on HDD, yes.</li><li>My RAM modules have scrambled frequencies, expect better performance.</li></ul>Issues: For weird reasons, if iterations are above 100it/s, gen takes 8-14 minutes. Interrupt and start it again for normal speed.<hr />Best Practices:No advanced prompt needed, straight describe as if using Flux or SDXL.<ul><li>describe first frame, describe next action to perform+details+adjectives, repeat, repeat; quality tags, negative tags</li></ul><ul><li>[the scene despicts a scared anime girl, she turns to face the camera, she closes her legs and touches her own chest; the camera moves farther away, highres, absurdres, detailed movements]</li></ul><ul><li>[the scene despicts a anime girl wearing a blindfold, she removes the blindfold, holding it close to her chest, her eyes are empty; no camera movement, highres, absurdres, detailed movements]</li></ul><ul><li>[a 3d model of chibi anime girl is holding two guns, she points the gun around while her hair sways with the wind, she looks at the camera]</li></ul>Showcase: <a target="_blank" rel="ugc" href="https://civarchive.com/posts/24529277/edit">https://civarchive.com/posts/24529277</a><hr />

Files with SHA256: 6de616d05e7b3f7299b84e7a168e3a5a25d1e268a8d335798b9269f65487d3c4