Skip to content

[OVEP] Fix UnsupportedOpMode logic for Reshape #703

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: ovep-develop
Choose a base branch
from

Conversation

mdvoretc-intel
Copy link

Assumptions about number of inputs to Reshape nodes were causing crashes. The check has been adjusted to be functional with Reshape only having one input.

@mdvoretc-intel mdvoretc-intel force-pushed the reshape_support_check_fix branch from 664e75b to c8cada8 Compare June 10, 2025 14:56
@sfatimar sfatimar requested a review from MayureshV1 June 12, 2025 03:03
@sfatimar
Copy link

Please update branch

Assumptions about number of inputs to Reshape nodes were causing crashes. The
check has been adjusted to be functional with Reshape only having one input.
@mdvoretc-intel mdvoretc-intel force-pushed the reshape_support_check_fix branch from c8cada8 to a3e1c1c Compare June 12, 2025 08:02
@@ -419,7 +419,8 @@ void DataOps::populate_op_mode_supported() {
UnsupportedOpMode obj = {{V_2023_1, V_2023_2, V_2023_3, V_2024_0, V_2024_1, V_2024_2,
V_2024_3, V_2024_4, V_2024_5, V_2024_6, V_2025_0, V_2025_1},
[this](const Node* node, const InitializedTensorSet&) {
const auto& input_arg = node->InputDefs()[1];
const auto& input_args = node->InputDefs();
const auto& input_arg = (input_args.size() > 1) ? input_args[1] : input_args[0];

Choose a reason for hiding this comment

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

May I ask how do we know the index 0 or 1 in the input_args is the parameter we want?

Copy link
Author

Choose a reason for hiding this comment

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

The original code assumes that the two provided parameters are the tensor containing the target shape and the data to be reordered. This change addresses the cases where the shape information is embedded as a metaparameter.

Choose a reason for hiding this comment

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

Then is it better to figure out why some cases put the shape information invisible or some not? So we can make it consistent for all cases. Just a brainstorming here, not a requirement for this PR.

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.

3 participants