Skip to content
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

[tools] new api and support dynamic models in python benchmark app #8582

Merged

Conversation

AlexeyLebedev1
Copy link
Contributor

@AlexeyLebedev1 AlexeyLebedev1 commented Nov 14, 2021

Tickets:

  • 68569

Details:

  • rewrite benchmark app with API 2.0.
  • support dynamic models.

Support dynamic models includes:

  • -shape parameter can be dynamic
  • -data_shape parameter represents input data shape
  • specifying -data_shape parameter isn't required in case image input, images will be processed with original shapes
  • -batch_size can be dynamic
  • -pcseq tool reports performance for each shape (group of shapes in case multi input network)
  • -legacy_mode - enable legacy scenario with inputs filling only once before measurements, default for static case
  • -full mode - include inputs filling into loop measurement, default for dynamic models.
  • Max/Avg/Min latency for both cases - dynamic and static, median latency is valid for static case only

Some changes in benchmark app:

  • Number of processed input files doesn't depend on number of infer requests anymore. Max number of images that is multiple number of data shapes will be used.
  • User able to pass path to directory in input file mapping.

nosovmik and others added 30 commits September 6, 2021 21:18
Includes API definition for trivial mean/scale operations (which don't require layout)

Mean/scale with 'layout' support will be done under separate task together
 with Layout

Current test code coverage: 100%
@AlexeyLebedev1 AlexeyLebedev1 requested a review from a team November 28, 2021 17:24
@akuporos akuporos merged commit f9bd740 into openvinotoolkit:master Dec 2, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
category: Core OpenVINO Core (aka ngraph) category: inference OpenVINO Runtime library - Inference category: tools OpenVINO C++ / Python tools
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet