Issue regarding building SparkNet for existing Spark Cluster

[prakhar21]

Hey, I am facing problem in building SparkNet, I have followed steps written under building it for existing Spark Cluster and since I don't plan to use GPU to have skipped CUDA step (Step 1). After I run 'sbt assembly', the error thrown is

[info] Run completed in 10 seconds, 779 milliseconds.
[info] Total number of tests run: 6
[info] Suites: completed 5, aborted 0
[info] Tests: succeeded 6, failed 0, canceled 0, ignored 9, pending 0
[info] All tests passed.
[error] Error during tests:
[error]     TensorFlowNetSpec
[error] (test:test) sbt.TestsFailedException: Tests unsuccessful

Is there any dependency that is not mentioned in the README.md ?

[robertnishihara]

The current build.sbt file in the repository pulls jars that were built with CUDA. In order to use only CPUs, change the line

resolvers += "javacpp" at "http://www.eecs.berkeley.edu/~rkn/snapshot-2016-03-05/"

in build.sbt to use the jars that were built for CPU usage ("http://www.eecs.berkeley.edu/~rkn/snapshot-2016-03-16-CPU/").

[prakhar21]

@robertnishihara Got that working, by making changes in the build.sbt
Thanks.

[robertnishihara]

Excellent!