mamba
mamba copied to clipboard
Fix stride handling for optional dt_bias and D
Hi, I found a bug for the stride values when dt_bias or D is None. When dt_bias or D is None, the stride should be (0, 0) rather 0.