onnx-tensorrt
onnx-tensorrt copied to clipboard
IshuffleLayer: when I change the shape of the tensor, the tensor's data becomes 0
TensorRT Version: 7.0
I want to reape the tensor of shape (1,1024,1) as (1,1024) , I used the following code:

However, the following exception occurred during inferencing:

What can I do to solve this problem, thank you!
TRT 7.0 is quite old, is this issue reproducible if you use the latest TRT version?