search

tensorfreitas / Siamese-Networks-for-One-Shot-Learning

Implementation of Siamese Neural Networks for One-shot Image Recognition
605 stars 181 forks source link

pretrained model #8

Closed mzarvandi closed 3 years ago

mzarvandi commented 5 years ago

Hi, I tried to train the network with a server using two GPUs still it's so much time consuming. Could you please share the weights and biases for pretrained network.

Thank you

tensorfreitas commented 5 years ago

Hi Maedeh,

I do not have the pre-trained model weights to share, but it should take some hours to train the model.

Best Regards Tiago Freitas

mzarvandi commented 5 years ago

HI Tiago,

I see. But I also configured that before 18 thousands epochs it doesn't reach to any better accuracy so it stops training, and by the way the best accuracy during test it reaches is 0.6125. I was wondering if you could tell me how to reach to a better accuracy. I also tried Adam optimizer it did not go further than about 70 percent accuracy. I will be thankful if you could give me a hint.

Regards,

On Mon, Sep 30, 2019 at 11:49 AM Tiago Freitas notifications@github.com wrote:

Hi Maedeh,

I do not have the pre-trained model weights to share, but it should take some hours to train the model.

Best Regards Tiago Freitas

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Goldesel23/Siamese-Networks-for-One-Shot-Learning/issues/8?email_source=notifications&email_token=AEZEW4NHQQH3RSNIDL6GYATQMGZC3A5CNFSM4I3OHHMKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD742UDY#issuecomment-536455695, or mute the thread https://github.com/notifications/unsubscribe-auth/AEZEW4POCATFI56I4N3SRA3QMGZC3ANCNFSM4I3OHHMA .

tensorfreitas commented 5 years ago

Hi Maedeh,

Sorry for the late response,

Are you training on Omniglot dataset with the exact same code/architecture?