stefan-matcovici comments

Results 5 comments of


                                            stefan-matcovici

Printing attack adversaries in a verbose mode, for better debugging purpose

After running with CUDA_LAUNCH_BLOCKING=1, I might get a more specific stacktrace: ``` Traceback (most recent call last): File "adversarial_test_model.py", line 56, in advdata = adversary.perturb(clndata, target) File "/nfshomes/smatcovi/works/pilot-project/ppenv/lib/python3.7/site-packages/advertorch/attacks/carlini_wagner.py", line 207,...

Printing attack adversaries in a verbose mode, for better debugging purpose

I might have sent you on the wrong track. Sorry about that. I was calling the attack with True as an argument for the num_of_classes. Maybe this is an issue...

[bug] apt update errors due to failing NVIDIA certificate verification

Having the same issue with images with torch

getting twitter.postTweet is not a function

Same issue here. That's what you suggested? ``` var Twitter = require('twitter-js-client').Twitter; var twitter = new Twitter(config); twitter.postTweet({status: "Test"}, error, success); ```

Cannot create a conda environment.

+ 1 same issue