The model of the target (gen) did not load correctly when the file was opened. (AudioPlayerStreamingMode)

Hello! I would like to ask the following questions.

Every time I open my file, there is a model error, as shown in the figure below.
(I am using Streaming mode, and there is no issue with Audio mode.)



However, as soon as I perform any of the following actions, it returns to normal, and the model returns to its original position:

  1. Play the default streaming audio file.
  2. Click on the CoreFullface node in LazyGraph (without making any changes, just clicking).
  3. Reset Network Name.

So, there is no problem in usage, but opening the file makes the screen look scary.
Is there a way to fix this? Or is this a bug?

After careful inspection, I found that CoreFullface has been correctly set and saved to Claire_v1. However, every time I reopen the file, it reverts to the default model shown in “Image 1.”


Therefore, my thoughts are as follows:

1. Is there a way to force a reload of Claire_v01 when opening the file?
2. Alternatively, is there a way to preload a short audio file?

I have to use streaming mode, and since this file will be shared with team members, we may also need to consider path settings.

In the latest releases, Audio2Face uses a generic head mesh as the driver rather than using a male or female mesh. The difference between this mesh and original Make or Claire networks is not applied to your mesh unless you play the animation or do something to trigger the deformer updates.

Here’s a script that can bake this difference into the generic mesh so it looks good when deformer is not activated:

import omni.usd
from omni.deform.shared.scripts.utils import get_final_deformed_points

# update this path to the a2f mesh
path = "/World/a2fTemplates/gen_xform_grp/gen"
stage = omni.usd.get_context().get_stage()
prim = stage.GetPrimAtPath(path)

thank you very much!!!

could you explain what is “bake this difference into the generic mesh”?
or trere is any doc about it?

thank you again

oh, I simply meant this script moves the vertices of the generic driver mesh to match the final mesh that was trained, e.g. Mark or Claire

where can I edit “generic mesh script”

thank you again!

You don’t need to change the script. You can run the script provided inside the script editor of Audio2Face which can be accessed from Window -> Script Editor

I fix that.
The path must be edit.
very thank you!



Hello, I would like to ask you again:
Apart from the head mesh, when opening the file, the character’s tongue is also in the wrong position.
Is there a similar script to fix this?

Thank you very much.

Can you change the path to point to the tongue, like this:

path = "/World/a2fTemplates/gen_xform_grp/c_bottomDenture_grp_hi/c_tongue_hi"

I use this script
it won’t work

import omni.usd
from omni.deform.shared.scripts.utils import get_final_deformed_points

path = “/World/a2fTemplates/gen_xform_grp/c_bottomDenture_grp_hi/c_tongue_hi”
stage = omni.usd.get_context().get_stage()
prim = stage.GetPrimAtPath(path)

He even caused the tongue to be stuck in place, not returning to the correct position, even when I started playing the voice, it didn’t return to normal.


Is it possible to share your usd file?


please check the link

please use “Chr_Yellow_Streaming.usd”
I have done the setting for streaming


It looks like your tongue path is here:
path = "/Yellow/a2fTemplates/gen_xform_grp/c_bottomDenture_grp_hi/c_tongue_hi"

But your stage is already corrupted with the wrong tongue points. Can you use this path on a stage you haven’t run the tongue script before?