Wan AI video model is probably the best open source option out there right now. In this guide we'll look at how to use image2video in Wan in ComfyUI. The workflow works in Swarm too.
Download free workflow here (at bottom of link) https://www.patreon.com/posts/123216177
1 - UPDATE YOUR COMFYUI OR SWARM
2 - Download Wan i2v Fp8 model https://huggingface.co/Kijai/WanVideo_comfy/tree/main
Wan2_1-I2V-14B-480P_fp8_e4m3fn.safetensors - /models/diffusion_models/
OR
Wan2_1-I2V-14B-720P_fp8_e4m3fn.safetensors - /models/diffusion_models/
3 COMFY ONLY - Download clip, text encoders and vae https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/tree/main/split_files
clip_vision_h.safetensors - /models/clip_vision/
umt5_xxl_fp8_e4m3fn_scaled.safetensors - /models/text_encoders/
wan_2.1_vae.safetensors - /models/vae/
OPTIONAL - If you have low vram, you can download GGUF models instead of FP8
480p https://huggingface.co/city96/Wan2.1-I2V-14B-480P-gguf/tree/main
720p https://huggingface.co/city96/Wan2.1-I2V-14B-720P-gguf/tree/main
Use Wan in the cloud and get 20% off TD Pro and +20% bonus on your first deposit at ThinkDiffusion (affiliate): https://www.thinkdiffusion.com/?via=sebastian
Chat with me in our community discord: https://discord.gg/dFB7zuXyFY
Stable Diffusion for Beginners Playlist https://www.youtube.com/playlist?list=PLXS4AwfYDUi5sbsxZmDQWxOQTml9Uqyd2
My Weekly AI Art Challenges https://www.youtube.com/playlist?list=PLXS4AwfYDUi7RvFm4K6lKBH_acaZQMKY4
My Stable diffusion workflow to Perfect Images https://youtu.be/4u-Ytioi3DM
ControlNet tutorial and install guide https://youtu.be/vFZgPyCJflE
Famous Scenes Remade by ControlNet AI https://youtu.be/wVbWZ-Ph9lE