Open HYER1N opened 1 year ago
I think that you can add a dictionary in the class AttentionStore to collect the cross-attention map for each step
and add the cross-attention map of each step to the dict in between_steps function
@wenqsun Thanks! so much helpful
In the paper, there is a visualization figure for each time step, but it is not in the code. If it's in the code, please let me know which part it is. If it's not in the code, please let me know which part I should check.