TastyPancakes/Flux.jl
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

update docstring

Browse Source
This commit is contained in:
Lyndon White 2019-03-11 18:40:29 -03:00
parent 025d9b678d
commit 7d247ea25b
1 changed files with 5 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
src/layers/basic.jl
View File

@ -130,7 +130,7 @@ end
Maxout(over) Maxout(over)
`Maxout` is a neural network layer, which has a number of internal layers, `Maxout` is a neural network layer, which has a number of internal layers,
which all have the same input, and the max out returns the elementwise maximium which all have the same input, and the maxout returns the elementwise maximium
of the internal layers' outputs. of the internal layers' outputs.
Maxout over linear dense layers satisfies the univeral approximation theorem. Maxout over linear dense layers satisfies the univeral approximation theorem.
@ -150,15 +150,16 @@ end
Maxout(f, n_alts, args...; kwargs...) Maxout(f, n_alts, args...; kwargs...)
Constructs a Maxout layer over `n_alts` instances of the layer given by `f`. Constructs a Maxout layer over `n_alts` instances of the layer given by `f`.
All other arguements (`args` & `kwargs`) are passed to the constructor `f`. The function takes no arguement and should return some callable layer.
Conventionally this is a linear dense layer.
For example the following example which For example the following example which
will construct a `Maxout` layer over 4 dense linear layers, will construct a `Maxout` layer over 4 internal dense linear layers,
each identical in structure (784 inputs, 128 outputs). each identical in structure (784 inputs, 128 outputs).
```julia ```julia
insize = 784 insize = 784
outsie = 128 outsie = 128
Maxout(Dense, 4, insize, outsize) Maxout(()->Dense(insize, outsize), 4)
``` ```
""" """
function Maxout(f, n_alts, args...; kwargs...) function Maxout(f, n_alts, args...; kwargs...)