Deeplearning4j: Nd4j/Libnd4j: CI issue with TF import argmax test

Created on 20 Aug 2018  路  5Comments  路  Source: eclipse/deeplearning4j

Test: "reductions/argmax3,4,5_-1" - passes locally on both of my machines (Windows 10, Kubunu 17.10). However is failing (timing out, 10 seconds - it should take <<10ms to execute) on multiple CI machines, for mac CPU, Linux AVX2, Linux AVX 512, Linux ppc64.

The other argmax tests (same shape, different axis) are OK.
Need to reproduce/isolate this before we can look into it more...

Bug DevOps

All 5 comments

Hypothetically #6209 fixes this issue.

Ah, cool, I missed that one :)

Hm, looks like no luck. Will keep looking.

Issue resolved

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.

Was this page helpful?
0 / 5 - 0 ratings