How to use it? You can watch my YouTube videos.
📺 Video Resources & Links
Qwen: Best-Ever Anime to Realistic (Z-image Redraw CFG 1.6) - Online Demo & Workflow: https://www.runninghub.ai/post/2002567462472302594/?inviteCode=sdhs0trb
Qwen: Best-Ever Anime to Real Life (Z-image | Official Scheduler Redraw): https://www.runninghub.ai/post/2002654471387770881/?inviteCode=sdhs0trb
One-Click Generation Link: https://www.runninghub.ai/ai-detail/2002658834701271042/?inviteCode=sdhs0trb
ModelScope Profile: https://modelscope.cn/profile/zhouwenbin1994
Model Download (Quark Drive):
https://huggingface.co/WarmBloodAban/Anything_to_Real_Characters
https://pan.quark.cn/s/1083abe30ee8
Rare Plugins Download: https://pan.quark.cn/s/6e5bce52eeda
AI Resources Cloud Drive (Models, Plugins, Tools, Trainers, etc.): https://zcn0otkmv1cq.feishu.cn/wiki/AgItw3W6pihFA2kzFpscyECmnzb?from=from_copylink
Test Images: https://pan.quark.cn/s/9701954775a4
⚠️ Hardware Requirement: NVIDIA GPU with 12GB VRAM or higher is required for local execution.
ComfyUI Anime to Human Integration Package (For first-time users of ComfyUI, you can download this, with a graphics memory of 12GB or more):
Description
FAQ
Comments (9)
I'm having this error with the flowmatch Euler node:
IndexError: index -1 is out of bounds for dimension 0 with size 0
What are the correct settings? I saw your video and copied the values you had
update: I fixed that but now the output in zimage group is only noise.
It should be that Flowmatch has updated the nodes, and some of the settings need to be modified. You can see the corresponding changes in my workflow
Where do I get 图像缩放V2_孤海 ?
search in google goohaitools github and install via git
The flowmatch node seems to have changed all values and positions after update which is why this might not work right away until u update the values. also , can you tell me what should be the values for start at step and end at step since these were not there in the previous version of the node and hence not in your video. thanks
using start step 0 and end step 8 did give a result. also had to connect v2 goohai tool node to the memory cleaner (force) node. however the anime pic is transforming in to an asian person even though the character doesn't have asian characteristics so not sure if it's because of the model since i used your anything to real model however the name inserted in the workflow previously was a bit different
also first generation is successful , second generation always crashes comfyui and restarting comfyui doesnt work, have to restart pc. only happens with this workflow. i have over 20 other workflows and no such issues there. i have a rtx 5090 64 gb ddr 5 ram, please fix it.
update: disabling z-image section fixed issue. plus i prefer image output of the first section over z-image anyways
Your flow looks amazing, but everything in chinese. I cant understand nothing. You can translate the flow please?