validation accuracy

[bzabdolahi]

Dear author
I'm using the BOSSbase_1.01 dataset, WOW algorithm 0.4bpp
4000 train, 1000 validation.
I train the model but I get valid_accuracy=0.5
could you help me what went wrong?

[EmadHelmi]

I also have this problem. Could you fix that?
I also train on the BOSSbase with 10000 images and I put 3000 for validation but exactly I get the same result as you mentioned.
I would be happy If you help me.

[bzabdolahi]

my problem was solved by using "DatasetPair" class for data loading. On Wednesday, February 12, 2020, 11:41:22 AM GMT+3:30, EmadHelmi <notifications@github.com> wrote: I also have this problem. Could you fix that? I also train on the BOSSbase with 10000 images and I put 3000 for validation but exactly I get the same result as you mentioned. I would be happy If you help me. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[EmadHelmi]

Can I have your code, please?
Or can I send my code to you?

[bzabdolahi]

I didn't change the code. I just set batch_size = 32 test_batch_size = 32 lr = 4e-1 use_batch_norm = False embed_otf = False

[TomTomTommi]

Sorry, I still not understand how to solve this problem, I try to use DatasetPair class for data loading, but the problem still here. coule you share the code, please?

[ByGary]

Does anyone sovle the above problem? The training loss and validation loss always stay at a high level.