I have a swin-transformer-tiny-based model which I have converted to ONNX format. The inference runs fine in python (I’m using onnxruntime) but importing the model in Unity gives me the error: TensorShape.ValueError: operands could not be broadcast together with shapes (729), (384,384).
My guess is that the inferred shape (729) is wrong and it should be (729,384)
Any input/help will be much appreciated. Thanks!
1 post - 1 participant