fixes thread safety in CustomObjectDetection not working #418
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
This change removes the thread_safe flag because it opts to always make object detection thread safe through the use of tf Sessions. It makes sense to remove the thread_safe flag since we need to setup the Session & Graph beforehand and don't want to introduce a bunch of extra thread_safe checks for no good reason. More details on the fix here: tensorflow/tensorflow#28287 (comment)
This is to ensure we get compatible versions of the two libraries. At the time of writing, if we let keras install 2.3.1 then we also get a bug which breaks threaded detection. See here for more details: keras-team/keras#13353
fixes keras version typo.
Codecov Report
@@ Coverage Diff @@
## master #418 +/- ##
==========================================
+ Coverage 49.96% 50.08% +0.11%
==========================================
Files 63 63
Lines 5856 5864 +8
==========================================
+ Hits 2926 2937 +11
+ Misses 2930 2927 -3 |
|
Any feedback on this @OlafenwaMoses or @rola93? |
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Outdated suggestions cannot be applied.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.
This change removes the thread_safe flag and instead opts to always make object detection thread safe through the use of tf Sessions. It makes sense to remove the thread_safe flag since we need to setup the Session & Graph beforehand in loadModel() and don't want to introduce a bunch of extra thread_safe flag checks. I added a unit test which runs the custom object detector on 3 threads as well.
More details on the fix here: tensorflow/tensorflow#28287 (comment)
I also locked the tensorflow and keras version down because theres a few bugs in later keras & tensorflow versions that will break threaded detection if tensorflow >2.0 and keras > 2.3.0 are installed. See here: keras-team/keras#13353