Sébastien Jalliffier Verne

Hi @carter54, I don't see anything obvious by reading the code. `smooth_step` also looks fine as it seems consistent with [OpenGL documentation](https://www.khronos.org/registry/OpenGL-Refpages/gl4/html/smoothstep.xhtml). It might help if you could post the...

This is supposed to be a warning. The demo handles this case and should work fine except that SIMD isn't supported by the browser. As a result, the default backend...

I'm not sure to understand what is not working then. In the screenshot you shared, we can see the background showing up at 60fps and given that no face is...

Thanks you for taking the time to share this screenshot. The console output is actually interesting. On my Safari 15.2 (no idea if the minor version does matter) where I...

Actually I tried to make every promise of the code fail and I didn't manage to reproduce both the face not showing and the background being rendered. So I'm now...

Thanks for your answer, I'm gonna investigate differently then.

Related issue: https://github.com/spasimir21/camera-processor/issues/5 That's interesting cause it seems that for now, @spasimir21 repo doesn't include yet all the WebGL stuff so I guess I can exclude everything related to this...

Please could you share the report from `chrome://gpu`? I'm running chrome v91.0.4472.101 without issue so maybe could you also try to upgrade it just in case.

Hi @carter54, The change you described looks fine to me. When stress testing a lot the app by changing a lots of times the model to load, I have in...

Maybe one thing that you could try though is to add a `printf` **after** loading the model, so after this block: https://github.com/volcomix/virtual-background/blob/main/tflite/tflite.cc#L67-L70 Maybe the memory issue doesn't happen when loading...