Model accuracy

[meirashaf]

Would you please provide the model evaluation metrics?

I tried to train the model using custom dataset, but the highest accuracy only reach 56.9% (much lower than what is stated in the original paper). Do you have any idea why it happen? Is it because of the custom dataset or pytorch adaptation?

It would be very helpful if you would provide metric accurary of the model you trained using this model pytorch adaptation. Thank you.