Tiling Spatial Concatenation for OpenVINO IR #2052
Merged
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
github-actions
bot
added
ALGO
eugene123tw
changed the title
eugene123tw
requested review from
goodsong81,
sungmanc,
jaegukhyun,
supersoob and
eunwoosh
sungmanc
reviewed
Apr 26, 2023
otx/algorithms/detection/configs/instance_segmentation/configuration.yaml
Outdated
Show resolved
Hide resolved
tests/unit/algorithms/detection/tiling/test_tiling_detection.py
Outdated
Show resolved
Hide resolved
goodsong81
requested changes
Apr 26, 2023
goodsong81
approved these changes
Apr 27, 2023
jaegukhyun
approved these changes
Apr 27, 2023
goodsong81
added
ENHANCE
goodsong81
pushed a commit
to goodsong81/training_extensions
that referenced
this pull request
May 2, 2023
* draft * udpate pytest skip tag * add spatial concatenation * code quality fixes * fix tests * add unit test for ir scale factor * revert pytest skip and address test failure in another PR * refactor variable name
goodsong81
added a commit
that referenced
this pull request
May 3, 2023
* Upgrade mmdeploy==0.14.0 from official PyPI (#2047) * Bug fix: value of validation variable is changed after auto decrease batch size (#2053) * Integrate new ignored loss in semantic segmentation (#2065) * Remove unused modules in semantic segmentation (#2068) * Add doc for fast data loading (#2069) * Bug fix: set gpu_ids properly (#2071) * Bug fix: Progress goes 100% and back 0 % repeatedly during auto decrease bs in Geti (#2074) * Fix tiling 0 stride issue in parameter adapter (#2078) * Update instance-segmentation tutorial documentation (#2082) * Optimize YOLOX data pipeline and add unit test for get_subset of Datu… (#2075) * Tiling Spatial Concatenation for OpenVINO IR (#2052) * Add spatial concatenation to deployment demo (#2089) --------- Signed-off-by: Songki Choi <[email protected]> Co-authored-by: Eunwoo Shin <[email protected]> Co-authored-by: Soobee Lee <[email protected]> Co-authored-by: Inhyuk Cho <[email protected]> Co-authored-by: Eugene Liu <[email protected]> Co-authored-by: Harim Kang <[email protected]> Co-authored-by: Jaeguk Hyun <[email protected]>
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
To optimize the inference process of an IR model, adjusting tiling parameters can be effective. For example, during training, 16x16 tiles with 100x100 pixels may be used. However, during inference, 8x8 tiles with 200x200 pixels can be used after reshaping the IR model once. This change can result in a 1/4 reduction in the number of calls, with expected internal parallelization in the OpenVINO runtime.
How to test
Checklist
License
Feel free to contact the maintainers if that's a concern.