MAHMOUDI Issam

it's a memory error, try decreasing the batch size for testing

@PoonamZ if even with batch_size_test=1 you still get the same error, this means that a whole image can't fit in your GPU memory , have you tried to do prediction...

@bundelesneha05 how much gpu memory do you have ? you need at least 2GB to fit that model

why do you mean by model does not converge?

Are you using Brats2017 datasets ? which numpy version do you use?

it's really weird in my local machine i can't reproduce the error can you give more details ?

yes it's probably because of memory when you transform large numpy arrays into np.float you'll quickly run out of memory , i've done this project on a machine that has...

@agungfathur check the filepath of the dataset

for everyone's asking for BraTS 2017 or BraTS 2018 datasets you can request them via this website : [https://ipp.cbica.upenn.edu/](url)

Yeah this would be great. It's the only reason preventing me from using the manager as of now. Having the ability to import Keystores would be great.