This workflow is for use with Kijai's Wanwrapper nodes and the new Master Model Wan 14b checkpoint merge that allows for text to video in high quality in only 1.5 minutes on a 4090 before upscaling.Master model download: <a target="_blank" rel="ugc" href="https://civarchive.com/models/1651125/wan2114bt2v-fusionx">https://civarchive.com/models/1651125/wan2114bt2v-fusionx</a>GGUFs of the above are available here:<a target="_blank" rel="ugc" href="https://huggingface.co/lym00/Wan14BT2V_MoviiGen_AccVid_CausVid_MasterModel_GGUF">https://huggingface.co/lym00/Wan14BT2V_MoviiGen_AccVid_CausVid_MasterModel_GGUF</a>Kijai's Wan Wrapper nodes:<a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-WanVideoWrapper">https://github.com/kijai/ComfyUI-WanVideoWrapper</a>The upscale model I'm using in the workflow:<a target="_blank" rel="ugc" href="https://openmodeldb.info/models/4x-NomosWebPhoto-RealPLKSR">https://openmodeldb.info/models/4x-NomosWebPhoto-RealPLKSR</a>1280x720p: On a 3090/4090, I'd change these settings for 720p: Wanvideo Empty Embeds changed to 1280x720. Wanvideo Block Swap node changed from 0 to 20. Wanvideo Sampler Node to 15 steps, cfg 1, shift 8, with euler/beta scheduler. You can probably get rid of the upscale node, but you should set the CRF quality setting on the video combine node from 30 to 19 so you don't compress out the increase in resolution.

Wanvideo 14b MasterModel Workflow

This is for the text to video version.

Files with SHA256: 33d12e1364061de7ea5cfa2b3fc78ceadffd79835f97bdec7474344c3cacb210