Input and Output Tensor Alignment - 1.4.1 English

Vitis AI RNN User Guide (UG1563)

Document ID
UG1563
Release Date
2021-12-03
Version
1.4.1 English

If the frame size does not match the aligned size (W), you need to zero-pad each frame. Similarly, the output dimension also has an aligned frame size. You can extract the required data if the actual frame size is less than the aligned size.