[Official] 🐞 Bug Report & Technical Support

Open Beta AI render
1

Thank you for participating in the AI Render Open Beta! Since this is a beta release, we highly appreciate your help in identifying bugs, technical issues, or usability problems.

This thread is your go-to place for:

  • Reporting bugs
  • Asking for technical support
  • Getting troubleshooting tips
  • Viewing known issues and upcoming fixes

:clipboard: How to Report a Bug Effectively

To help us resolve your issue quickly, please provide the following details:

:wrench: Plugin Version

  • The version of AI Render you are currently using.

:computer: System Specs

(You can find this information in the Reallusion Hub under “System Info”)

  • CPU, GPU, RAM, Operating System

:gear: ComfyUI Version (if applicable)

  • Only include this if you’ve updated or modified ComfyUI yourself.

:memo: Detailed Description of the Problem

  • What were you trying to do?
  • What did you expect to happen?
  • What actually happened?

:receipt: Error Messages

  • Include screenshots of any error pop-ups or messages in the console, if available.

:repeat: Steps to Reproduce

  • Provide a step-by-step walkthrough so we can replicate the issue on our end.

:file_folder: Example Files

  • Include your history folder and any sample outputs if possible.
  • If you’re comfortable sharing, your project file or ComfyUI workflow would also be extremely helpful.
3

What happens if I already have ComfyUI installed on my PC? I recently installed the latest version of ComfyUI via Git. I imagine both instances can coexist side by side, but it seems redundant to me.

4

you still have to do the same thing and add the patch and UI for iclone. It loads its all into iclone. just install like the video or transcript says. you install 2 things. Good luck with it.

5

i am using the latest Icone 8 and CC4. When I load my character and then click on AI render, i do not get the setting window. it does not pop up

6

Hi MBusch

To ensure a smooth experience for users who are new to ComfyUI, we recommend using the version of ComfyUI we provide during the current beta phase, as it’s ready to use with our workflows.

If you have custom workflow requirements and prefer not to maintain two separate ComfyUI environments, you can switch to LAN Server mode via the Settings panel when working with your own setup. This will allow the plugin to connect directly to your custom ComfyUI instance.

You can refer to the following step-by-step guide to create a custom workflow.

7

Hi mackemmy,

After launching iClone 8 or Character Creator 4, you can go to the top menu and select Script > Console Log to check for any error messages.

If no message appears initially but you can see the AI Render toolbar, try clicking on any AI Render function and then check the Console Log again.

Please take a screenshot of any error messages that appear and share it with us. We’ll help you troubleshoot the issue.

8

for some reason, it started working. not sure what happend

9

Just a couple of notes.

  • Firing up Preview resets a custom frames range set on timeline. It resets to default - All frames.

  • When using a Video mode and hitting Preview, it starts videos export right away and then when it finishes, it does AI preview of the current frame. There is NO way to test random frames preview in Video mode. I wish we had a separate button for video export. I know I could use image mode for that, but it’s inefficient to keep all settings in sync and jumping between tabs.

  • I think someone already mentioned about the ability to save an entire configuration to a file along with prompts and all other settings and load them at will.

  • AI does not seem to recognize Humanoid characters (even those build with AccuRig). Armature never detected.

10

Hi 4u2ges,

Thank you for the feedback. We’ll take it into consideration and discuss how to make adjustments.

Regarding the first issue you mentioned about resetting the frame range. Just to confirm, are you clicking Preview after setting the Start/End range in the timeline?

11

Hi Melvin,

Basically yes, It actually does not matter at which point the range was set. I can open an existing project with preset range and it would reset as soon as I hit Preview.
BUT it seems only happening in Image mode.

Also in Video mode if I change the the range within AI, that would change it on timeline as well. But not backward. Changing range on timeline is reflected on iClone render panel but not within AI.

12

Thanks a lot for the extra input. We’ll keep it in mind and make some adjustments.

To clarify the second point about the video mode preview:
The current video generation setup is based on how Wan 2.1 works. It uses the first frame as a reference to guide the whole video, while the rest of the frames are mainly driven by the control data.

Because of this, previewing any random frame in the middle of the video may give results that don’t fully match what the final video will look like. The preview is mainly meant to help you confirm the overall style and direction before rendering.

But we’re still actively exploring what’s possible with ComfyUI, and there might be ways to improve or expand this behavior in the future.

If you want to avoid re-rendering the whole video every time just to check the style, you can turn off the “Always Render New” option. This will reuse your last output and skip the export step.

Hope this is helpful.

1 Like
13

Hi, keep getting this error: RuntimeError: Error(s) in loading state_dict for Resampler:
size mismatch for proj_in.weight: copying a param with shape torch.Size([768, 1280]) from checkpoint, the shape in current model is torch.Size([768, 1664]).

14

Hi stownsendd,

May I ask what resolution you’re currently using for output? We’ve noticed that some specific dimensions can cause AI model to misread the data.

15

Hi Melvin,

Thank you for your reply!

I changed the output 1024 x 1024 to viewport resolution and it worked!, I will try some others when I get back from work and let you know the results.

16

After setting it to the viewport resolution, I can now scale it in the AI render tab and all is working as it should now…thank you for your help Melvin, much appreciated!

17

I have downloaded but not installed the plugin yet -just avoiding unnecessary waste of disk space as these files are BIG ! I would like to know first "Can it render as 16x9 or other ? " From the [ intro ]videos Iv’e seen the render format is 1x1. Otherwise I don’t see the use for movies if usual film formats not possible. Thanks

18

Stable Diffusion 1.5 is trained on 512x512.

Wan 2.1 Fun Control says in its description that it supports multiple resolutions: 512, 768, 1024

Wan 2.1 Vace supports 480p for the 1.3B model while the 14B model supports 480p and 720p.

If you use other models, for example Stable Diffusion XL, then you can use 1024x1024, if you use Flux you can use resolutions from 0.1 to 2.0 megapixel.

So there is no general answer for all. You need to experiment because the models usually accept small variations to image sizes. But if your image is way too small or too large you will get artifacts or the image creation will brake down completely.

1 Like
19

If this was in response to my post about render formats - then thankyou

20

Ok installed everything per video, i got a bunch of errors in the cmd log, does the installer install git? I got the green light even with the errors but would like to resolve them… says git error will persist until a path defined…yet i dont see git as an installed app, suggestions? anyone else get these errors?

21

Was Git added to the Windows System PATH variable? The command line will look there where to find the git.exe otherwise it doesn’t know where it is located.

Here you’ll find screenshots with a step-by-step guide. In my experience, it’s enough the add the bin folder containing the git.exe to the Path and not necessary to add the exe itself.

https://linuxhint.com/add-git-to-path-windows/