minibatchqueue or arrayDatastore drops my data precision from double to single

Question

Runcong Kuang 2022-9-28

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1813485-minibatchqueue-or-arraydatastore-drops-my-data-precision-from-double-to-single

评论： Runcong Kuang 2022-9-29

在 MATLAB Online 中打开

I get XTrain from MNIST by processImagesMNIST and put it on GPU, so its type is gpuArray dlarray.

then I use these code to make minibatches:

```

miniBatchSize = 128;
dsTrain = arrayDatastore(XTrain,IterationDimension=4);
% numOutputs = 1;
mbqTest = minibatchqueue(dsTrain,1, ...
    MiniBatchSize = miniBatchSize, ...
    MiniBatchFcn=@preprocessMiniBatch, ...
    MiniBatchFormat="SSCB", ...
    PartialMiniBatch="discard");
% numObservationsTrain = size(XTrain,4);
% numIterationsPerEpoch = ceil(numObservationsTrain / miniBatchSize);
% numIterations = numEpochs * numIterationsPerEpoch;
%% test batch order
i=0;
while hasdata(mbqTest)
    i = i+1;
    x = next(mbqTest);
    if ~hasdata(mbqTest)
        disp(i)
    end
end
```

And I find that x is single gpuArray dlarray, XTrain is gpuArray dlarray.

I wonder which part makes it lower the precision.

And how to avoid this?

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Runcong Kuang 2022-9-29

I find that minibatchqueue has a 'name,value' pair which is outputCast, 'single' as default. So I assign 'double' to it. But I haven't checked if the GPU training process does use single precision, as mentioned by @Walter Roberson.

But thanks for your input @Walter Roberson.

请先登录，再进行评论。

请先登录，再回答此问题。