Setup for Run_TCAV.ipynb fails
Line 126 in imagenet_and_broden_fetcher.py (bytes = urllib.request.urlopen(all_images)) fails. This seems to be a result of bad URLs (e.g. http://imagenet.stanford.edu/api/text/imagenet.synset.geturls?wnid=n02391049).
Thanks for raising this issue! It seems that Imagenet has now closed url-based download, and only allow downloads for those how signed-in to their website. https://image-net.org/download.php We'll have to look at how else we can provide the examples (since we can't simply copy their data and store locally--that would be violation of the copyright). Please stay tuned.
Hello, BeenKim. have your team got a new example (maybe new datasets) for ‘tcav’ running checking?