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

better alternative to basemodel

Browse Source
This commit is contained in:
Mike J Innes 2017-02-28 16:42:48 +00:00
parent 5f1f2ebaa2
commit 4d4979b401
3 changed files with 5 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
docs/src/examples/char-rnn.md
View File

@ -16,15 +16,13 @@ N = length(alphabet)
Xs, Ys = getbatches(input, alphabet), getbatches(input[2:end], alphabet) Xs, Ys = getbatches(input, alphabet), getbatches(input[2:end], alphabet)
basemodel = Chain( model = Chain(
Input(N), Input(N),
LSTM(N, 256), LSTM(N, 256),
LSTM(256, 256), LSTM(256, 256),
Affine(256, N), Affine(256, N),
softmax) softmax)
model = Chain(basemodel, softmax)
m = tf(unroll(model, nunroll)) m = tf(unroll(model, nunroll))
@time Flux.train!(m, Xs, Ys, η = 0.1, epoch = 1) @time Flux.train!(m, Xs, Ys, η = 0.1, epoch = 1)
@ -38,5 +36,5 @@ function sample(model, n, temp = 1)
return string(s...) return string(s...)
end end
sample(basemodel, 100) sample(model[1:end-1], 100)
``` ```

6
examples/char-rnn.jl
View File

@ -13,15 +13,13 @@ N = length(alphabet)
Xs, Ys = getbatches(input, alphabet), getbatches(input[2:end], alphabet) Xs, Ys = getbatches(input, alphabet), getbatches(input[2:end], alphabet)
basemodel = Chain( model = Chain(
Input(N), Input(N),
LSTM(N, 256), LSTM(N, 256),
LSTM(256, 256), LSTM(256, 256),
Affine(256, N), Affine(256, N),
softmax) softmax)
model = Chain(basemodel, softmax)
m = tf(unroll(model, nunroll)) m = tf(unroll(model, nunroll))
@time Flux.train!(m, Xs, Ys, η = 0.1, epoch = 1) @time Flux.train!(m, Xs, Ys, η = 0.1, epoch = 1)
@ -35,4 +33,4 @@ function sample(model, n, temp = 1)
return string(s...) return string(s...)
end end
sample(basemodel, 100) sample(model[1:end-1], 100)

2
src/layers/chain.jl
View File

@ -31,4 +31,4 @@ graph(s::Chain) =
shape(c::Chain, in) = c.shape shape(c::Chain, in) = c.shape
Base.getindex(c::Chain, i::AbstractArray) = Chain(c.layers[i]) Base.getindex(c::Chain, i::AbstractArray) = Chain(c.layers[i]...)