Why do I get the error"Training using trainNetwork failed. Duplicate table variable name: 'input'"

When I use ’combine‘ to connet 'TrainA_Train' and 'TrainB_Train' , then I get the error"Training using trainNetwork failed. Duplicate table variable name: 'input'"

Regarding ’TranA‘, it is a file read using the 'imageDatastore'.

Regarding ’TranB‘, it is also a file read using the 'imageDatastore'.

this is my code:

TrainA = imageDatastore("C:\Users\Administrator\Desktop\data1","IncludeSubfolders",true,"LabelSource","foldernames");

[TrainA_Train, TrainA_Valid, TrainA_test] = splitEachLabel(imdsTraingray,0.6,0.2,"randomized");

TrainB = imageDatastore("C:\Users\Administrator\Desktop\data2","IncludeSubfolders",true,"LabelSource","foldernames");

[TrainB_Train, TrainB_Valid, TrainB_test] = splitEachLabel(imdsTraingray,0.6,0.2,"randomized");

Train = combine(TrainA_Train,TrainB_Train);

Valid = combine(TrainA_Valid,TrainB_Valid);

opts = trainingOptions("adam",...

"ExecutionEnvironment","auto",...

"InitialLearnRate",0.0001,...

"LearnRateDropFactor",0.01,...

"LearnRateDropPeriod",10,...

"LearnRateSchedule","piecewise",...

"MaxEpochs",20,...

"MiniBatchSize",40,...

"Shuffle","every-epoch",...

"Plots","training-progress",...

"ValidationData",Valid);

[net, traininfo] = trainNetwork(Train,lgraph,opts);

Note:I omitted the network model

I want to train a deep neural network with 2 inputs and 1 output,but I don't konw how to revise this problem.

Answer 1

[Cris LaPierre]

I think this is because you are incorrectly using the same datastore (imdsTraingray?) for your splitEachLabel commands. I think you should be using TrainA and TrainB.

TrainA = imageDatastore("C:\Users\Administrator\Desktop\data1","IncludeSubfolders",true,"LabelSource","foldernames");

[TrainA_Train, TrainA_Valid, TrainA_test] = splitEachLabel(imdsTraingray,0.6,0.2,"randomized");

TrainB = imageDatastore("C:\Users\Administrator\Desktop\data2","IncludeSubfolders",true,"LabelSource","foldernames");

[TrainB_Train, TrainB_Valid, TrainB_test] = splitEachLabel(imdsTraingray,0.6,0.2,"randomized");

Try the following instead

TrainA = imageDatastore("C:\Users\Administrator\Desktop\data1","IncludeSubfolders",true,"LabelSource","foldernames");
[TrainA_Train, TrainA_Valid, TrainA_test] = splitEachLabel(TrainA,0.6,0.2,"randomized");
TrainB = imageDatastore("C:\Users\Administrator\Desktop\data2","IncludeSubfolders",true,"LabelSource","foldernames");
[TrainB_Train, TrainB_Valid, TrainB_test] = splitEachLabel(TrainB,0.6,0.2,"randomized");