leezu commented on issue #11509: Hanging flaky test test_operator.test_norm @
Python 3: GPU Win
URL:
https://github.com/apache/incubator-mxnet/issues/11509#issuecomment-408991191
@marcoabreu is it feasible to restart the CI Pipeline with the same global
seed (not only test seed)? Jenkins
leezu commented on issue #11509: Hanging flaky test test_operator.test_norm @
Python 3: GPU Win
URL:
https://github.com/apache/incubator-mxnet/issues/11509#issuecomment-408985161
I can't reproduce this on p3 and p2 instances so far. I keep running this
test in a loop for now. Do the
leezu commented on issue #11509: Hanging flaky test test_operator.test_norm @
Python 3: GPU Win
URL:
https://github.com/apache/incubator-mxnet/issues/11509#issuecomment-408977295
I'm looking into the issue now.
This is an
leezu commented on issue #11509: Hanging flaky test test_operator.test_norm @
Python 3: GPU Win
URL:
https://github.com/apache/incubator-mxnet/issues/11509#issuecomment-404907122
@anirudh2290 @KellenSunderland Only the part that caused assertion errors
during numerical checking of the
leezu commented on issue #11509: Hanging flaky test test_operator.test_norm @
Python 3: GPU Win
URL:
https://github.com/apache/incubator-mxnet/issues/11509#issuecomment-402893579
The latter error could be due to that numpy.linalg.norm is not numerically
stable (neither is the mxnet one).