OVMS 2022.3.0.1 release #1721
Merged
OVMS 2022.3.0.1 release #1721
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
Without this we sporadically hit the PIPELINE_NOT_LOADED_YET status in CI. This can hapen under heavy machine workload as the timeout for waiting was 10ms and sporadically it was ~70ms on some machines. Adding this status to allowed statuses list for this test would lower the test value since we can hit it when eg. there is unsuccessful validation. If the threshold is not increased the PIPELINE_NOT_LOADED_YET status code would encourage to try to send request again so it seems that's better to wait a bit longer than force client to retry with request. JIRA:CVS-63360
* fixes for tests * fixes for tests 2
- wait for all inferences to be finished before checking sequence manager count - increase to max 5s timeout for slower machines to pass - lower number of inferences from 100 to 25
* Fixes for 2 issues: CVS-62467 - All ovms-c tests have the same start time CVS-86713 - Remove CPU_THROUGHPUT_STREAMS from python tests.
* Makefile targets * Tests off * Ccache Co-authored-by: Damian Kalinowski <damian.kalinowski@intel.com>
…e and custom gpu driver installation (#1637)
- update links in PyPi overview - update base image for RedHat Dockerfile to UBI 8.7 - remove upper bounds for dependencies - make adjustments so it can work with >= 1.24 numpy version and unify behavior for pre and post 1.24 for inhomogeneous array creation attempt - add Makefile argument for also testing with numpy==1.23 - change location of TF protos to stay under ovmsclient namespace - that way it can reside in the same Python environment with tensorflow without overwriting one another
CVS-99305
* Fix AUTO_CASUSES_COPY * Fix big parameter passed by value * Fix nullptr check in cliparser * Remove unreachable code * Add virtual destructor to base class * Narrow down active time request metric for model in DAG * Check for return value from library in s3 filesystem * Remove uninitialized c-tor field * Fix unsigned check JIRA:CVS-100296
* C-API DAG support implementation * Unify getting servable name & info about native file format for DAG * Rename native file format conversion function convertBinaryRequestTensorToOVTensor -> convertNativeFileFormatRequestTensorToOVTensor * Move DAGS implementation into separate directory * Divide Makefile steps * Increased threshold for cppclean warnings due to: - adding additional static logger - cppclean not understanding template class instatiation in cpp file and taking it for forward declaration (EntryNode, ExitNode, ExitNodeSession) *Increase coverage JIRA:CVS-95627 Co-authored-by: Damian Kalinowski <damian.kalinowski@intel.com>
sync with main
* remove catalogs from previous build * change minimal protobuf compiler version for building ovmsclient
* Fix checking for cpu count on system Using std::hardware_concurrency could return 0. JIRA:CVS-99839
…1687) There was issue when model is reloaded or added back to config.json. mapping_config.json file was not applied for first version in set of version to load due to wrong path to mapping_config.json.
Co-authored-by: Trawinski, Dariusz <dariusz.trawinski@intel.com>
- fix memory leak in example cpp client (loaded image memory was not freed) - updated doc with code snippets accordingly - added error handling to KServe cpp async client - fix TFS cpp client issue with accurracy being rounded to integer - fixed azure credential object copy
#1660) removed duplication in a dockerfile
* add support for session token parameter in s3 buckets * added documentation changes
Error will be reported when metrics are configured both in the CLI and the config file.
While validation of such requests worked fine the issue was that during reloading incorrectly original input name was used which resulted in model effectively not reloading with new shape and failure during request deserialization. JIRA:CVS-103489
Fix parsing issue for data with different level of arrays
…onfig, shape, layout (#1700) * Fix schema for model_version_policy * Add checks for plugin config & unified model mapping schema * Add test for model mapping * Add tests for model_version_policy * Improve schema for layout, shape and add tests JIRA:CVS-102870
- fix issue with negative value REST shape causing abnormal gRPC proto creation - fix issue with gRPC negative shape value, causing abnormal behavior in post-validation steps - added unit tests - added support for signed values in shape in test framework for better corner case testing
dkalinowski
approved these changes
Feb 27, 2023
dtrawins
approved these changes
Feb 27, 2023
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.
No description provided.