jurijsnazarovs / bayesian_nn Goto Github PK

How we can use custom dataset with your model..I tried to implement but always getting after 1st epoch? can you guide me please

Hello, I am trying to run your code on my pc which has a 32GB RAM CPU and a 3GB GTX 1060 GPU. And I am using my own dataset with 100 * 7 images in total. However, I got 'Out Of Memory' error before the very first epoch for net like vgg11, even if the batch size is 1.
With CPU, the error is

RuntimeError: $ Torch: not enough memory: you tried to allocate 0GB. Buy new RAM! at ..\aten\src\TH\THGeneral.cpp:204

With GPU, the error is

RuntimeError: CUDA error: out of memory

I wonder if it is a hardware problem, or is there some restriction for the input size of images for different NNs? If the first case, what will be the basic hardware requirement?

Thank you very much.