Kent Gauen

Thank you @madfff. I appreciate your comment. Using your script, I have 39.71/0.958 rather than 39.74/0.958. Still, I find it strange we are missing [39.89db - 39.74db =] 0.15 db....

I am bumping this so it doesn't close from inactivity. I think being unable to reproduce the key result is worth investigating further. I am excited to hear from the...

Thank you for the reply! My validation results match your reported validation results at 39.77 db.

I am not sure I understand why this comment is closed. I am able to match your validation results. I am unable to match your test results.

Maybe this was an accident -- I can reopen a new issue no problem!

Ah I didn't understand all these results are on the validation set. Thank you for the pointer here -- I am unable to find this detail in your paper.

This is defined in the base class _LRScheduler (c.f. [see here](https://github.com/pytorch/pytorch/blob/30fccc53a9a1ae88db39ef7dd4de35bd2c174ca4/torch/optim/lr_scheduler.py#L43)). The base_lrs by default takes the "lr" parameter set by the optimizer (c.f. [see here](https://github.com/pytorch/pytorch/blob/30fccc53a9a1ae88db39ef7dd4de35bd2c174ca4/torch/optim/optimizer.py#L235-L240)). I believe the "lr"...

see issue #10 this can be fixed by pasting the following code snippet into the `script/rundissect_pytorch.sh` ``` # For expanding globs http://stackoverflow.com/questions/2937407 function globexists { set +f test -e "$1"...

The "globexists" function should be defined in the file. I am guessing the authors read [this stack overflow page](https://stackoverflow.com/questions/2937407/test-whether-a-glob-matches-any-files), and forgot to include this in their bash script. If you...

Yes, this github repo can run "network dissection" on any deep learning model if the files are provided properly to the code. You will first have to get VGG16 trained...