When we load glb, if the model has skeletal animation, the loading will be very complicated. I have an idea: Can we serialize/deserialize the parsed Mesh object instead of re-parsing and generating the Mesh object every time we ImportMesh?
No matter how we cache the glb file loading results, it is still very time-consuming to generate a Mesh object using glb data each time. Do we have any solution to optimize this logic?
Could you share a repro ?