Skip to content
This repository was archived by the owner on Jul 1, 2023. It is now read-only.

Adding conv transpose 1d & 3d #174

Merged
merged 20 commits into from
Nov 19, 2019
Merged

Adding conv transpose 1d & 3d #174

merged 20 commits into from
Nov 19, 2019

Conversation

Shashi456
Copy link
Contributor

I'll add the tests. but till then could someone check the logic once and tell me if i'm wrong or right?

@rxwei rxwei requested a review from jekbradbury June 4, 2019 18:03
@Shashi456 Shashi456 changed the title Adding conv transpose 3d Adding conv transpose 1d & 3d Jun 5, 2019
@Shashi456 Shashi456 changed the title Adding conv transpose 1d & 3d Adding conv transpose 3d Jun 5, 2019
@Shashi456 Shashi456 changed the title Adding conv transpose 3d Adding conv transpose 1d & 3d Jun 5, 2019
@Shashi456
Copy link
Contributor Author

Shashi456 commented Jun 5, 2019

I've added 1d transpose as well, this is something not found in the keras library. But could someone check the logic of that as well? I'll add the rest of the conv layers throughout the week.
I'll start writing test cases now

@Shashi456
Copy link
Contributor Author

Note: The build passes.

@Shashi456
Copy link
Contributor Author

@rxwei Any updates on taking a look at this?

@rxwei rxwei requested review from rxwei and dan-zheng and removed request for jekbradbury June 14, 2019 06:03
Copy link
Contributor

@rxwei rxwei left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the ping. Could you please add a test for each of these?

@Shashi456
Copy link
Contributor Author

@rxwei There seems to some disconnect between the tensorflow/Keras Convtranspose APIs and our APIs, could you help me write some tests for this?

@rxwei
Copy link
Contributor

rxwei commented Jun 18, 2019

@rxwei There seems to some disconnect between the tensorflow/Keras Convtranspose APIs and our APIs, could you help me write some tests for this?

We will try to take a look later this week!

@Shashi456
Copy link
Contributor Author

Shashi456 commented Aug 19, 2019

@rxwei Updated this PR, the build and tests pass on the latest nightly.

Sorry for the delay!

@Shashi456
Copy link
Contributor Author

done

@Shashi456
Copy link
Contributor Author

@rxwei could you approve this

@noDerivative public let strides: (Int, Int, Int)
/// The padding algorithm for convolution.
@noDerivative public let padding: Padding
@noDerivative public let paddingIndex: Int
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Add a doc comment for this property.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@rxwei this property isn't really exposed to the users, it's a temporary property which allows us to handle computation based on what kind of padding it is. we currently support .same and .valid.
Do we still document this?

@noDerivative public let stride: Int
/// The padding algorithm for convolution.
@noDerivative public let padding: Padding
@noDerivative public let paddingIndex: Int
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Add a doc comment for this property.

@burmako
Copy link

burmako commented Nov 14, 2019

@Shashi456 I've been assigned to help move this PR over the finish line. At a glance, things are looking great :)

If you could resolve the conflicts and add documentation comments for paddingIndex (something that explains why those fields are necessary), I'll run the tests and then we'll merge. Thank you!

@Shashi456
Copy link
Contributor Author

@burmako this property isn't really exposed to the users, it's a temporary property which allows us to handle computation based on what kind of padding it is. we currently support .same and .valid.
Do we still document this?
When we support the third padding feature, i think this will be omitted.

@burmako
Copy link

burmako commented Nov 14, 2019

@Shashi456 I think a comment along the following lines would be great:

/// Temporary property which allows us to handle computation based on padding.
/// Will be removed when issue #1234 is fixed.

@Shashi456
Copy link
Contributor Author

@burmako does this work?

@burmako
Copy link

burmako commented Nov 19, 2019

Thank you! I'll be merging now :)

@burmako burmako merged commit 2d06188 into tensorflow:master Nov 19, 2019
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants