Skip to content

Fixed Dropout inference issue in OpenVINO ONNX conversion #29234

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 2 commits into
base: master
Choose a base branch
from

Conversation

thesakshidiggikar
Copy link

🛠 Fix Dropout Inference Issue in OpenVINO ONNX Conversion

🔍 Issue:

While converting an ONNX model with Dropout layers, inference results were incorrect due to improper handling of Dropout during conversion. This issue is reported in #21615.

🔧 Fix Implemented:

  • Updated the ONNX model conversion to handle Dropout layers correctly.
  • Ensured that inference results are consistent with expected outputs.
  • Tested with multiple models to verify accuracy.

✅ Testing & Verification:

  • Ran inference using OpenVINO’s inference_engine and verified outputs.
  • Compared results with native ONNX inference to confirm correctness.
  • Passed all unit tests in model_conversion module.

📌 Related Issue:

Fixes #21615


Please review and let me know if any changes are needed! 🚀

@thesakshidiggikar thesakshidiggikar requested review from a team as code owners March 3, 2025 03:49
@thesakshidiggikar thesakshidiggikar requested review from ilya-lavrenov and removed request for a team March 3, 2025 03:50
@sys-openvino-ci sys-openvino-ci added the ExternalPR External contributor label Mar 3, 2025
@p-wysocki
Copy link
Contributor

@gkrivor could you please take a look?

@p-wysocki p-wysocki requested a review from gkrivor March 5, 2025 13:17
@mlukasze
Copy link
Contributor

ping @gkrivor
@thesakshidiggikar code style needs to be fixed at first, I suppose

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

[Bug]: Port for tensor name x was not found
4 participants