Giters

dmlc / MXNet.jl

MXNet Julia Package - flexible and efficient deep learning in Julia

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Question: Minimal Tutorial

abieler opened this issue · comments

commented

I am not smart enough to extrapolate from the MNIST example what to
do with data and labels that come in form of regular Julia arrays.
Say, X = Matrix{Float}(1000, 10) and y = Vector{Int}(1000)

Is there a minimal example for a 3 layer perceptron applied to this data?

Thanks
Andre

commented

There is a simple regression example that shows how to deal with providing data from arrays in https://github.com/dmlc/MXNet.jl/blob/master/examples/regression-example.jl

commented

Thanks! I somehow managed to not see this example...

commented

I created a Jupyter Notebook MLP tutorial taking MNIST data from Kaggle and accessing it using an ArrayDataProvider. You can see it here: https://github.com/ultradian/julia_notebooks/blob/master/mxnet/mnistMLP.ipynb

I would love to contribute to the MXNet.jl project with further documentation. How would I best do this?

commented

@ultradian Thanks a lot! I suggest one of the two options, whichever you prefer that might be easier for maintenance:

  • Make a PR to put your jupyter notebook to the MXNet.jl repo. Under some directory such as examples/jupyter-notebooks.
  • Keep your notebook hosted on your side, but add a link to it somewhere in the doc.
commented

@pluskid Thank you for all your work. I'm going for option 2, and just submitted a PR #227. Hopefully, I'll be able to contribute something more substantial in the future!