assert error %2!=0

[123liluky]

I saw assert error in closed issues, Thank you

[chuckyee]

It means that at the model is attempting to do a 2x2 max-pool on a feature map that isn't even in the number of pixels. Either check the size of the inputs or reduce the number of layers in the model.

[bayonettablack]

I don't understand author means,how to solve it?

[chuckyee]

Hi @superfinale, I've updated the defaults.config file -- this shouldn't be an issue anymore.

The problem is caused by attempting to downsample (via max-pool) an image with a dimension not divisible by 2. For example 216x256 -> 108x128 -> 54x64 -> 27x32. If you try to downsample again, the code will throw an error.

[bayonettablack]

@chuckyee I try it again,but it still has the same bug...

I want to know how to process the image data and then train my own data,thank you!

[chuckyee]

What is the input dimension of the images you're using and what is depth set to?

[bayonettablack]

I haven't train my own data,Now the dataset is you used,I want to reproduce the result and then train my own data.Was the route incorrect?

[bayonettablack]

@chuckyee

[makman7]

@lumialengine run with padding="same" or default.config there shouldn't be any problem.