ml-commons
ml-commons copied to clipboard
ml-commons provides a set of common machine learning algorithms, e.g. k-means, or linear regression, to help developers build ML related features within OpenSearch.
### Description This PR is for the option 1 in the issue #2151 ### Issues Resolved [List any issues this PR will resolve] ### Check List - [ ] New...
### Description We want to use a MLModelTool to firstly do tool selection and then limit the tool root agent can use. It will improve the tool selection accuracy. To...
Disk circuit breaker ([code link](https://github.com/opensearch-project/ml-commons/blob/2.x/plugin/src/main/java/org/opensearch/ml/breaker/DiskCircuitBreaker.java#L21)) has a hard coded threshold as 5 GB. We should make it as a dynamic setting.
The integration test failed at distribution level for component ml-commonsVersion: 2.12.0Distribution: tarArchitecture: arm64Platform: linuxPlease check the logs: https://build.ci.opensearch.org/job/integ-test/6700/display/redirect * Test-report manifest:* - https://ci.opensearch.org/ci/dbc/integ-test/2.12.0/9061/linux/arm64/tar/test-results/6700/integ-test/test-report.yml _Note: Steps to reproduce, additional logs and...
**What is the bug?** This might not be a bug in the code or IT test itself. Flaky tests are caused by the testing environment which sometimes triggered java.net.ConnectException and...
### Describe the bug **Issue Description:** After successfully deploying a machine learning model in AWS OpenSearch using the POST request to `/_plugins/_ml/models/WgtRZY0BjrbwPvlSkd2A/_deploy`, the model initially shows as deployed when checked...
**What is the bug?** When we add a field to the system index’s mapping (e.g. ml-model, ml-connector, etc), we should update the mapping proactively in the backend if the mapping...
We found that the error status code not popped out correctly. We fixed the bug in https://github.com/opensearch-project/ml-commons/pull/1443/files We should add more integ test to cover this case.
**Is your feature request related to a problem?** We don't have auto benchmarking worklfow for models for now. **What solution would you like?** We can build some auto benchmarking tools...