Open AeroDEmi opened 1 year ago
I Have the same issue for TorchScript, any idea?
Hello, I have the same problem with ONNX. I obtain an onnx file where the inference works with only one resolution but I get an error with reshapes sizes with all the other input resolutions... Did you manage to find a solution to your problem ?
has annyone gotten this to work ?
Is there a way to convert this model to onnxruntime? I have been trying with no luck so far, so many errors with
Wh, Ww = (H + 1) // 2, (W + 1) // 2
and lots of trace warningsTracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect.