Skip to content

Conversation

@lattner
Copy link
Contributor

@lattner lattner commented Jul 17, 2018

This will start producing graph_op in more cases, and inches towards strict deabstraction being on by default.

This will mostly be about updating testcases, which I'll iterate in the PR.

@lattner lattner added the tensorflow This is for "tensorflow" branch PRs. label Jul 17, 2018
@lattner
Copy link
Contributor Author

lattner commented Jul 17, 2018

@swift-ci please test tensorflow

@lattner
Copy link
Contributor Author

lattner commented Jul 17, 2018

I expect this to fail a bunch of testcases, I'll iterate here in the PR and request review when I converge.

file now blocking the other issues from being reported in partitioning, so move that test to the right file.
@lattner
Copy link
Contributor Author

lattner commented Jul 17, 2018

@swift-ci please test tensorflow

1 similar comment
@lattner
Copy link
Contributor Author

lattner commented Jul 17, 2018

@swift-ci please test tensorflow

@lattner lattner requested review from mhong and rxwei July 17, 2018 23:09
@lattner
Copy link
Contributor Author

lattner commented Jul 17, 2018

@swift-ci please test tensorflow

2 similar comments
@lattner
Copy link
Contributor Author

lattner commented Jul 17, 2018

@swift-ci please test tensorflow

@lattner
Copy link
Contributor Author

lattner commented Jul 17, 2018

@swift-ci please test tensorflow


public func shapeError() {
// expected-error @+1 {{tensor literal should have 9 scalars for this shape, but has 8}}
let _ = Tensor<Float>(shape: [1, 3, 3, 1],
Copy link

Choose a reason for hiding this comment

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

the 9 in "tensor literal should have 9 scalars for this shape" could be a bit confusing -- do we want to give more context on what the extra scalar is -- is that the meta type?

@lattner lattner merged commit 71b6da3 into tensorflow Jul 17, 2018
@lattner lattner deleted the const-tensors branch July 17, 2018 23:32
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

tensorflow This is for "tensorflow" branch PRs.

4 participants