meshed-memory-transformer
meshed-memory-transformer copied to clipboard
Reproduce the Visualization of Attention States
Hello!
Thanks for this great code!
I find there is the visualization of attention states for captions, like this.
I want to reproduce the visualization results, but I cannot find the corresponding code in this repo.
I read the paper (Axiomatic attribution for deep networks) that you cited, and I think it's not easy for me to reproduce the visualization results correctly.
May I have the code to produce the visualization results, or is there anything I missed?
Thank you very much!
I also want to do it but fails. have you reproduced the visualization results successfully? Thanks
has the same question