☄️I suggest using the new Lightspeed checkpoint: <a target="_blank" rel="ugc" href="https://civarchive.com/models/1802623">https://civarchive.com/models/1802623</a>⬆️⬆️⬆️⬆️⬆️⬆️⬆️⬆️⬆️⬆️⬆️⬆️⬆️⬆️⬆️⬆️⬆️⬆️ GGUF Quants of Wan 2.1 for lower VRAM usage.I suggest to use TeaCache for ~2x speedup if it is working. Info: Teacache may be broken ATM and can produce deformed videos. !!!All samples are 640x640 with TeaCache enabled and a Init-Image.Source: Source (gguf): <a target="_blank" rel="ugc" href="https://huggingface.co/city96/Wan2.1-I2V-14B-480P-gguf/tree/main">https://huggingface.co/city96/Wan2.1-I2V-14B-480P-gguf/tree/main</a> from <a target="_blank" rel="ugc" href="https://huggingface.co/city96">city96</a>Here are the results of some testing ...VRAM and RAM calculations:<ul><li>Your VRAM should be 1~2 GB higher as the model sizee.g. 16GB (VRAM): Model 12GB -2GB = 10 -&gt; good to go</li><li>You should always spare 1-2 GB for you OS to operate functional</li><li>Your RAM should at least be 16, better 32 GB</li><li>For Q8 you need 40+ GB RAM</li></ul>Hints for faster but reasonable good results I2V (Image to Video):<ul><li>Use a low resolution inside the maximum pixel count</li><li>Upscale the complete video after generation with a tool of your liking</li><li>Use 20-30 steps, 20 is good, 30 needs significant more time but preserves details<ul><li>The step increase is linear so, 10 more steps results from e.g. 10 min to 15 min</li></ul></li><li>Use CGF 4-6 (6 are good most of the time), 4 are sometimes better if lighting goes off</li><li>The image as Init-Image you use should have the same ratio as the video resolution</li></ul>Good resolutions:9:16/ 16:9 with 480x832 / 832x480 = 480p3:4 / 4:3 with 480x640 / 640x480 = 480p<hr /><span style="background-color:rgb(26, 27, 30);color:rgb(141, 141, 143);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">YOU are responsible for outputs as always! If you make <a target="_blank" rel="ugc" href="https://civarchive.com/content/tos"><span style="background-color:rgb(26, 27, 30);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">ToS</a><span style="background-color:rgb(26, 27, 30);color:rgb(141, 141, 143);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px"> violating content and I get aware I WILL report this.

Wan-AI / Wan2.1 Video Model (Safetensors) - GGUF Quants - i2v 14B

wanAIWan21VideoModel_wan21I2v14b480pQ80.gguf

wan2.1-i2v-14b-480p-Q8_0

wan2.1-i2v-14b-480p-Q6_K

wan2.1-i2v-14b-480p-Q5_K

wan2.1-i2v-14b-480p-Q3_K

ON-THE-FLY 实时生成！Wan-AI 万相/ Wan2.1 Video Model (multi-specs) - CausVid&Comfy&Kijai - workflow included

<a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org">Comfy-Org</a>/<a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged">Wan_2.1_ComfyUI_repackaged</a>Wan 2.1 repackaged for ComfyUI use. For examples see: <a target="_blank" rel="ugc" href="https://comfyanonymous.github.io/ComfyUI_examples/wan">https://comfyanonymous.github.io/ComfyUI_examples/wan</a>---Wan 2.1 Models<a target="_blank" rel="ugc" href="https://github.com/Wan-Video/Wan2.1">Wan 2.1</a> is a family of video models.---Files to DownloadYou will first need:Text encoder and VAE:<a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/tree/main/split_files/text_encoders">umt5_xxl_fp8_e4m3fn_scaled.safetensors</a> goes in: ComfyUI/models/text_encoders/<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">wan_2.1_vae.safetensors</a> goes in: ComfyUI/models/vae/---Video Modelsfiles go in: ComfyUI/models/diffusion_models/These examples use the 16 bit files but you can use the fp8 ones instead if you don’t have enough memory.---Workflows---Text to Video<a target="_blank" rel="ugc" href="https://comfyanonymous.github.io/ComfyUI_examples/wan/text_to_video_wan.json">Workflow in Json format</a>files (put it in: ComfyUI/models/diffusion_models/). You can also use it with the 14B model.---Image to Video<a target="_blank" rel="ugc" href="https://comfyanonymous.github.io/ComfyUI_examples/wan/image_to_video_wan_example.json">Workflow in Json format</a>Note this example only generates 33 frames at 512x512 because I wanted it to be accessible, the model can do more than that. The 720p model is pretty good if you have the hardware/patience to run it.

Files with SHA256: db961e33570a3a7902614ea52b6e77a33d284dfc85743a05e880a02da1164b45