Skip to content

Conversation

@RandySheriffH
Copy link
Contributor

Restrict the use of cuda FusedConv to be only on floats.

@RandySheriffH RandySheriffH requested a review from a team as a code owner February 11, 2021 08:03
Copy link
Contributor

@oliviajain oliviajain left a comment

Choose a reason for hiding this comment

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

Tested and confirmed that change resolves error for earlier failing model. Testing others.

ONNX_NAMESPACE::TensorProto_DataType_FLOAT) {
continue;
}
if (graph_utils::IsSupportedOptypeVersionAndDomain(next_node, "Relu", {6, 13})) {
Copy link
Contributor

Choose a reason for hiding this comment

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

FusedConv only supports float on CPU as well so this could be moved up

@RandySheriffH RandySheriffH merged commit a07a14d into master Feb 12, 2021
@RandySheriffH RandySheriffH deleted the ExcludeTypeFromFuzed branch February 12, 2021 21:30
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants