-
Notifications
You must be signed in to change notification settings - Fork 444
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
Fix detection performance degradation #3691
Merged
sungchul2
merged 1 commit into
openvinotoolkit:releases/2.1.0
from
sungchul2:fix-det-perf
Jun 28, 2024
Merged
Fix detection performance degradation #3691
sungchul2
merged 1 commit into
openvinotoolkit:releases/2.1.0
from
sungchul2:fix-det-perf
Jun 28, 2024
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…to affect training
sungchul2
requested review from
samet-akcay,
harimkang,
vinnamkim,
eugene123tw,
kprokofi,
chuneuny-emily,
sovrasov,
GalyaZalesskaya,
negvet,
goodsong81,
yunchu,
wonjuleee and
eunwoosh
as code owners
June 28, 2024 01:20
harimkang
approved these changes
Jun 28, 2024
wonjuleee
approved these changes
Jun 28, 2024
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
wonjuleee
pushed a commit
that referenced
this pull request
Jun 29, 2024
* [hotfix] Update for fix workflow issues (#3668) * Update README & CODEOWNERS (#3659) * Update Engine's docstring & CLI --help outputs (#3658) Update Engine's CLI docstring & HelpFormatter * Fix unit test for semantic segmentation to run it without mmseg (#3670) fix test to run wo mm * Enable ruff & ruff-format into otx/algo/classification/backbones (#3667) * Enable ruff in otx.algo.classification.backbones * Fix unit-test * Fix docstring * Disable integration test in pr-merge workflow (#3677) * Add TV MaskRCNN Tile Recipe (#3655) * add tv maskrcnn recipe * add unit test * add tile size * fix bug * Align integration test to exportable code interface update for release branch (#3676) * align integration test to exportable code update * add error message * align vp expected output name * Refactor exporter for anomaly task and fix a bug with exportable code (#3672) * refactor exporter * remove *to * reply comments * resolve problem with local temp file * Fix NNCF MaskRCNN-Eff accuracy drop (#3680) * fix nncf maskrcnn eff accuracy drop * update ignored scope * fix typo * Update pandas version constraint (#3679) * Include more models to export test into test_otx_e2e (#3678) * enable export test * re-enable failed models * exclude exportable code test from anomaly task * apply it to e2e test code * align with pre-commit * Fix optimize with Semi-SL data pipeline (#3684) Disable unlabeled dataset with optimize * Fix MaskRCNN SwinT NNCF Accuracy Drop (#3685) fix swin-t nncf kinda * Bump MAPI version (#3686) * Add rotated det OV recipe (#3687) * Add rotated det ov recipe * fix e2e test cli * Move assigning tasks to Models from Engine to Anomaly Model Classes (#3683) * Move Task assign into Model with Anomaly Task * Fix openvino model class * Refactoring detection modules (#3636) * Organize common functions that are located in each det model into OTXDetectionModel - __init__, _build_model, _create_model, _customize_inputs, _customize_outputs, get_classification_layers * Create common directory for modules used across tasks * Remove DictConfig to use dict instead * Update docstring * Refactoring - export related things - grouped importing backbones, necks, heads, losses - move assigners, coders, prior_generators, samplers which are located in heads to utils * Fix segmentation fault on VPM PTQ (#3689) Add forcing `num_workers` to 0 * Fix detection performance degradation (#3691) Update to apply filtering to only `test_step` and `predict_step` not to affect training * Bump datumaro to 1.7.0 (#3669) * bump datumaro to 1.7.0 * apply changes on datumaro 1.7.0 * Add maskrcnn_r50_tv_tile (#3692) * Update version string and fill missing changelog (#3693) --------- Co-authored-by: Harim Kang <harim.kang@intel.com> Co-authored-by: Prokofiev Kirill <kirill.prokofiev@intel.com> Co-authored-by: Eugene Liu <eugene.liu@intel.com> Co-authored-by: Eunwoo Shin <eunwoo.shin@intel.com> Co-authored-by: Vladislav Sovrasov <sovrasov.vlad@gmail.com> Co-authored-by: Kim, Sungchul <sungchul.kim@intel.com> Co-authored-by: Emily Chun <emily.chun@intel.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
ticket no. : 144952
Through #3634, detection predictions were fixed to be filtered.
But, these filtered predictions affect training performance.
This PR includes:
After this fix, performance is recovered.
The below table shows only
atss_mobilenetv2
, but performance tables for all models are in the ticket.How to test
Checklist
License
Feel free to contact the maintainers if that's a concern.