Fix generate square subsequent mask + attribute readers for convolutional layers

ankane / torch.rb

Deep learning for Ruby, powered by LibTorch

Other

704 stars 30 forks source link

Fix generate square subsequent mask + attribute readers for convolutional layers #30

Closed orlando-labs closed 2 years ago

orlando-labs commented 2 years ago

Hi, @ankane. Please review given fix and update. The update addresses the issue when output shape precalculus is needed. And that involves reading the attributes of the layers. Thanks

ankane commented 2 years ago

Hey @orlando-labs, thanks for another PR. Can you remove the changes to module.rb and module_test.rb?

orlando-labs commented 2 years ago

Oh, sure, done

ankane commented 2 years ago

Great, thanks!