-
Notifications
You must be signed in to change notification settings - Fork 254
/
test-wasserstein.jl
76 lines (63 loc) · 2.11 KB
/
test-wasserstein.jl
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
data_type = Float32
using Mocha
srand(123)
N = 1000
d = 4
lambda = 0.1
sinkhorn_iter = 10
n_hidden_layer = 2
n_hidden_unit = 64
test_wasserstein = true
X = rand(data_type, d, N)
Y_onehot = zeros(data_type, 1, N)
Y = zeros(data_type, d, N)
for i = 1:N
Y[indmax(X[:,i]),i] = 1
Y_onehot[i] = indmax(X[:,i])-1
end
# ground metric
M = zeros(d, d)
for i = 1:d
for j = 1:d
M[i,j] = abs(i-j)
end
end
backend = DefaultBackend()
init(backend)
data_layer = MemoryDataLayer(batch_size=50,
data=Array[X, test_wasserstein ? Y : Y_onehot])
hidden_layers = map(1:n_hidden_layer) do i
InnerProductLayer(name="ip$i", output_dim=n_hidden_unit, neuron=Neurons.ReLU(),
tops=[symbol("ip$i")], bottoms=[i == 1 ? :data : symbol("ip$(i-1)")])
end
pred_layer = InnerProductLayer(name="pred", output_dim=d,
tops=[:pred], bottoms=[symbol("ip$n_hidden_layer")])
if test_wasserstein
softmax_layer = SoftmaxLayer(bottoms=[:pred], tops=[:softpred])
loss_layer = WassersteinLossLayer(bottoms=[:softpred, :label],
M = M, lambda = lambda, sinkhorn_iter = sinkhorn_iter)
else
loss_layer = SoftmaxLossLayer(bottoms=[:pred, :label])
end
common_layers = [hidden_layers..., pred_layer]
if test_wasserstein
common_layers = [common_layers..., softmax_layer]
end
net = Net("Training", backend, [data_layer, common_layers..., loss_layer])
method = SGD()
params = make_solver_parameters(method, max_iter=10000,
lr_policy=LRPolicy.Fixed(0.01), mom_policy=MomPolicy.Fixed(0.9))
solver = Solver(method, params)
add_coffee_break(solver, TrainingSummary(), every_n_iter=100)
# for simplicity, we just use the training data as test...
data_layer_test = MemoryDataLayer(batch_size=100, data=Array[X, Y_onehot])
acc_layer = AccuracyLayer(bottoms=[:softpred, :label])
test_layers = [data_layer_test, common_layers..., acc_layer]
if !test_wasserstein
softmax_layer = SoftmaxLayer(bottoms=[:pred], tops=[:softpred])
test_layers = [test_layers..., softmax_layer]
end
test_net = Net("Testing", backend, test_layers)
add_coffee_break(solver, ValidationPerformance(test_net), every_n_iter=500)
solve(solver, net)
shutdown(backend)