DC-GAN tutorial based on TensorFlow

[Dsantra92]

Related to #107

[Dsantra92]

Oops! Spelling mistakes! 😬

[ToucheSir]

Thanks @Dsantra92! Just some minor changes from me below:

[logankilpatrick]

@Dsantra92 great work so far, if we can address the suggestions above, this should be good to go!

[Dsantra92]

@logankilpatrick I am done with the writing part, just refactoring the code a bit. I should be able to wrap it up by today.

[Dsantra92]

Anything more I should add?

[ToucheSir]

I don't think so, but let me have another look tomorrow.

[logankilpatrick]

Right now I get:

ERROR: UndefVarError: gen_ps not defined
Stacktrace:
 [1] train_generator!(gen::Chain{Tuple{Dense{typeof(identity), Matrix{Float32}, Flux.Zeros}, BatchNorm{typeof(relu), Vector{Float32}, Float32, Vector{Float32}}, var"#6#7", ConvTranspose{2, 4, typeof(identity), Array{Float32, 4}, Flux.Zeros}, BatchNorm{typeof(relu), Vector{Float32}, Float32, Vector{Float32}}, ConvTranspose{2, 4, typeof(identity), Array{Float32, 4}, Flux.Zeros}, BatchNorm{typeof(relu), Vector{Float32}, Float32, Vector{Float32}}, ConvTranspose{2, 4, typeof(tanh), Array{Float32, 4}, Flux.Zeros}}}, disc::Chain{Tuple{Conv{2, 4, typeof(identity), Array{Float32, 4}, Vector{Float32}}, var"#8#10", Dropout{Float64, Colon}, Conv{2, 4, typeof(identity), Array{Float32, 4}, Vector{Float32}}, var"#9#11", Dropout{Float64, Colon}, typeof(flatten), Dense{typeof(identity), Matrix{Float32}, Vector{Float32}}}}, fake_img::Array{Float32, 4}, opt::ADAM, ps::Zygote.Params, hparams::HyperParams)
   @ Main ./REPL[35]:3
 [2] train(hparams::HyperParams)
   @ Main ./REPL[39]:47
 [3] top-level scope
   @ REPL[41]:1
 [4] top-level scope
   @ ~/.julia/packages/CUDA/YpW0k/src/initialization.jl:52

[ToucheSir]

Side note: has anyone looked into writing tutorials as literate.jl or jmd files so that they can be easily verified and run?

[logankilpatrick]

Great work on this @Dsantra92 and thanks for the review @ToucheSir!

[Dsantra92]

Thanks @logankilpatrick 🙌. This was a great learning experience all thanks to you and @ToucheSir .