experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Downgrate the "missing shape at index" warning.

Browse Source 
This comes up because gradient code mutates function call operations. We'll already have (correct) shapes for the user-visible outputs, so in most cases this won't be too serious. We should fix the issue, but in the meantime there's nothing users can/need to do.

PiperOrigin-RevId: 304515374
Change-Id: I1ec132d216878454a205cf7bfc648b0582643ab8
This commit is contained in:
Allen Lavoie 2020-04-02 17:58:37 -07:00 committed by TensorFlower Gardener
parent 1ca0d81b0a
commit a51ef16118
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
tensorflow/core/common_runtime/shape_refiner.cc
View File

@ -85,8 +85,8 @@ Status InferShapesForFunctionSubNode(const Node* node, ShapeRefiner* refiner,
// TODO(b/134547156): TEMPORARY WORKAROUND. If input shape handle is not set
// in outer context, set _Arg node output shape to unknown.
if (outer_context->input(index).SameHandle(ShapeHandle())) {
LOG(WARNING) << "Function instantiation has undefined input shape at "
<< "index: " << index << " in the outer inference context.";
VLOG(1) << "Function instantiation has undefined input shape at "
<< "index: " << index << " in the outer inference context.";
node_context->set_output(0, node_context->UnknownShape());
} else {
node_context->set_output(0, outer_context->input(index));