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

[onert/odc] Auto-compilation. Input/output buffers in Execution #14383

Merged

Conversation

Torrero
Copy link
Contributor

@Torrero Torrero commented Nov 29, 2024

This PR for odc:auto-compilation introduces OdcInfo structure and input/output buffers in Execution.

For Issue.
From Draft.

ONE-DCO-1.0-Signed-off-by: Evgenii Maltsev [email protected]

@Torrero Torrero force-pushed the onert_odc_autocompilation_execution_changes branch from cf0fdb2 to d3f92a4 Compare November 29, 2024 18:09
@hseok-oh
Copy link
Contributor

hseok-oh commented Dec 2, 2024

Looks good for input/output buffer getter.
BTW, what is purpose of OdcInfo? Representation of model loading status? or Input model type of ODC?

@Torrero
Copy link
Contributor Author

Torrero commented Dec 2, 2024

Looks good for input/output buffer getter. BTW, what is purpose of OdcInfo? Representation of model loading status? or Input model type of ODC?

@hseok-oh Purpose of OdcInfo is for representation of model loading status (as for quantized as for compiled model). It's used in the run_with_auto_compilation method for providing loading status and checking the code generation status.

@hseok-oh
Copy link
Contributor

hseok-oh commented Dec 2, 2024

@hseok-oh Purpose of OdcInfo is for representation of model loading status (as for quantized as for compiled model). It's used in the run_with_auto_compilation method for providing loading status and checking the code generation status.

Then that seems independent with ODC. Is that used in ODC? Otherwise, it may be better to handled in session (in nnfw_api_internal), IMHO.

@Torrero
Copy link
Contributor Author

Torrero commented Dec 2, 2024

@hseok-oh Purpose of OdcInfo is for representation of model loading status (as for quantized as for compiled model). It's used in the run_with_auto_compilation method for providing loading status and checking the code generation status.

Then that seems independent with ODC. Is that used in ODC? Otherwise, it may be better to handled in session (in nnfw_api_internal), IMHO.

@hseok-oh you are right, this structure is not used in ODC, it's used in nnfw_session.
I deleted OdcInfo structure and added AutoCompilationState to nnwf_session in the draft and changed the usage in run_with_auto_compilation

…cution

This PR for `odc:auto-compilation` introduces OdcInfo structure and input/output buffers in Execution.
For [Issue]( Samsung#13288).
From [Draft](Samsung#13530).

ONE-DCO-1.0-Signed-off-by: Evgenii Maltsev [email protected]
@Torrero Torrero force-pushed the onert_odc_autocompilation_execution_changes branch from d3f92a4 to d6974be Compare December 2, 2024 17:10
@Torrero
Copy link
Contributor Author

Torrero commented Dec 2, 2024

Rebased this PR and deleted OdcInfo structure

Copy link
Contributor

@hseok-oh hseok-oh left a comment

Choose a reason for hiding this comment

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

LGTM

Copy link
Contributor

@stamalakhov stamalakhov left a comment

Choose a reason for hiding this comment

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

LGTM! Thank you!

@Torrero Torrero changed the title [onert/odc] Auto-compilation. OdcInfo and input/output buffers in Execution [onert/odc] Auto-compilation. Input/output buffers in Execution Dec 3, 2024
Copy link
Contributor

@SlavikMIPT SlavikMIPT left a comment

Choose a reason for hiding this comment

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

LGTM

@hseok-oh hseok-oh merged commit 240bda7 into Samsung:master Dec 4, 2024
9 checks passed
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.

4 participants