In[]:=
trainingData=​​{{97,3}{1,0,0,0,0,0,0,0,0,0},​​{90,10}{0,1,0,0,0,0,0,0,0,0},​​{82,18}{0,0,1,0,0,0,0,0,0,0},​​{71,29}{0,0,0,1,0,0,0,0,0,0},​​{46,54}{0,0,0,0,1,0,0,0,0,0},​​{38,62}{0,0,0,0,0,1,0,0,0,0},​​{26,74}{0,0,0,0,0,0,1,0,0,0},​​{15,85}{0,0,0,0,0,0,0,1,0,0},​​{3,97}{0,0,0,0,0,0,0,0,1,0},​​{0,100}{0,0,0,0,0,0,0,0,0,1}};
In[]:=
net=NetChain[{LinearLayer[50],ElementwiseLayer["ReLU"],LinearLayer[10],SoftmaxLayer[]}];
In[]:=
trainedNet=NetTrain[net,trainingData,MaxTrainingRounds->1000,LossFunction->CrossEntropyLossLayer["Binary"]]
Out[]=
NetChain
Inputport:
vector(size: 2)
Outputport:
vector(size: 10)

In[]:=
test={{97,3}};
In[]:=
trainedNet/@test
Out[]=
{{4.03943×
-22
10
,0.,1.,4.12993×
-25
10
,0.,0.,0.,0.,1.28386×
-13
10
,0.}}