apache / mxnet

Lightweight, Portable, Flexible Distributed/Mobile Deep Learning with Dynamic, Mutation-aware Dataflow Dep Scheduler; for Python, R, Julia, Scala, Go, Javascript and more
https://mxnet.apache.org
Apache License 2.0
20.77k stars 6.79k forks source link

test_nccl.test_nccl_pushpull was disabled #11707

Open szha opened 6 years ago

szha commented 6 years ago

The unit test in title have been using fixed seed to mask flakiness. Suggested action:

  1. Evaluate whether the test is flaky without fixed seed. If not, remove seed. Else move to 2
  2. If test is flaky, determine whether it's an actual uncaught edge case. If so, fix the operator. Else move to 3
  3. If numerical instability is inevitable, adjust tolerance level appropriately.
apeforest commented 6 years ago

Thanks for filing this issue. We will investigate this Flaky test

haojin2 commented 6 years ago

Can you change the title to "disabled test"? This is actually a disabled test

perdasilva commented 5 years ago

12477 seems to be handling it- so I've closed my PR (#12382)

perdasilva commented 5 years ago

@lebeg should this issue be closed?

lebeg commented 5 years ago

@perdasilva https://github.com/apache/incubator-mxnet/pull/12477 wasn't merged, so it should stay open.