raft-optical-flow

[yangsu2022]

Hi @raymondlo84 @eaidova
This is my new PR.
The old one: #966

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[paularamo]

@yangsu2022 Please rename this PR.

[ryanloney]

I tried running on Arc A770 dGPU and this is where it fails:

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

OpenVINO-dev-contest commented on 2023-05-19T06:06:13Z
----------------------------------------------------------------

Can you correct the ERROR message here ?

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

OpenVINO-dev-contest commented on 2023-05-19T06:06:14Z
----------------------------------------------------------------

Line #11.        :param W: width of model input

Please align the parameters with its definition: H->h

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

OpenVINO-dev-contest commented on 2023-05-19T06:06:15Z
----------------------------------------------------------------

can we use output_layer defined previous to get the output value ?

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

OpenVINO-dev-contest commented on 2023-05-19T06:06:16Z
----------------------------------------------------------------

Line #6.        RY = 15

Shall we add some comments on what these parameters means ?

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

OpenVINO-dev-contest commented on 2023-05-19T06:06:17Z
----------------------------------------------------------------

Pls fix the ERROR in log message

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

OpenVINO-dev-contest commented on 2023-05-19T06:06:18Z
----------------------------------------------------------------

Line #56.                Batch, Channel, height, width = compiled_model.input(0).shape

How about wrap these openvino's function in a Class, then we can reuse it in image and video test. You have already use N, C, H, W = input_layer.shape to get the input information before, shall we unify them ?

[yangsu2022]

I tried running on Arc A770 dGPU and this is where it fails:

Yes, OV22.3 not support 7D tensor on GPU, which fixed with OV23.0.
https://jira.devtools.intel.com/browse/CVS-100599?jql=text%20~%20%22raft%22

[ryanloney]

@yangsu2022 sorry, I was running in wrong env. I tried with pre-release 23.0 and it's running on GPU now.

BTW why is benchmark_app run twice in the notebook?

Here is what I'm seeing for Arc A770 dGPU benchmark:

Unfortunately this is what the visualization looks like when running on iGPU and dGPU:

I'm using 2023.0.0.dev20230427 from PyPI

[github-actions]

This PR will be closed in a week because of 2 weeks of no activity.

[github-actions]

This PR was closed because it has been stalled for 2 week with no activity.