Deeplearning4j: Type inconsistency

[hanbaoan123]

Issue Description

I am using rldj for some experiments, but I found type inconsistency between rl4j and nd4j. The return type of INDArray's shape() is long[]，but the parameter type Learning.makeShape() needed is int[].

Version Information

Rl4j and nd4j are all 1.0.0-SNAPSHOT.

Aha! Link: https://skymindai.aha.io/features/ND4J-43

[saudet]

Please feel free to update it to long[] and send a pull request! Thanks for reporting

[hanbaoan123]

Please feel free to update it to long[] and send a pull request! Thanks for reporting

In other words, this is a bug?

[saudet]

Yes, in the sense that we could in theory need to have dimensions larger than Integer.MAX_VALUE.

[hanbaoan123]

Yes, in the sense that we could in theory need to have dimensions larger than Integer.MAX_VALUE.

I got it and will send a pull request later.