Experimental support of Torch.jl

[findmyway]

We used to have support for Knet.jl in addition to Flux.jl, but it was dropped since ReinforcementLearning.jl@v0.4.0. The main reason was that Knet.jl is not very easy to extend. However, most functionalities have remained. I just notice that https://discourse.julialang.org/t/ann-announcing-torch-jl/42390 is relatively mature now. So maybe we can add an example here and compare the performance?

wait FluxML/Torch.jl#27

wait FluxML/Torch.jl#35

[oysteinsolheim]

Is anything changed with the new release of Knet 1.4?
https://discourse.julialang.org/t/ann-knet-1-4-0-accelerating-cuarrays/45206

[findmyway]

Thanks for sharing the news. I'll definitely check it out!