[SYSTEMDS-3681] Add stepGLM builtin function. - #2523
Conversation
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## main #2523 +/- ##
===========================================
Coverage 71.38% 71.39%
- Complexity 48753 50546 +1793
===========================================
Files 1571 1632 +61
Lines 188912 196238 +7326
Branches 37067 38201 +1134
===========================================
+ Hits 134857 140095 +5238
- Misses 43593 45129 +1536
- Partials 10462 11014 +552 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
christinadionysio
left a comment
There was a problem hiding this comment.
Thank you for you PR, @Bruno-Ignaz!
I have a few minor things that should be improved before we can push your code into the upstream. It would be appreciated if you could have a look.
|
Thanks @christinadionysio for the detailed feedback! |
|
@Bruno-Ignaz could you please look at the failing github action? After this is fixed, I will merge it in. |
Wire startLocalFedMonitoring through FederatedWorkerUtils.waitForWorker so the monitoring backend's port-bind is polled instead of slept on (fixes flaky FederatedCoordinatorIntegrationCRUDTest), migrate FederatedLogicalTest to the bulk startLocalFedWorkers(int[]) API, and drop the now-unused FED_WORKER_WAIT_S and FED_MONITOR_WAIT constants.
…ts (apache#2472) FedWorkerReadMatrixCompress.verifyRead failed roughly once per ten component-test CI runs because it called FederatedTestUtils.wait(1000) to give the worker time to finish its async compression (kicked off by CompressedMatrixBlockFactory.compressAsync), then asserted that the returned block was a CompressedMatrixBlock. On a contended runner the 1 s sleep was not enough, the subsequent read returned the still- uncompressed block, and the assertion failed. Surefire's rerunFailingTestsCount=2 hid this as a "Flake" rather than a job failure. Add FedWorkerBase.awaitCompressed(long id), which polls getMatrixBlock at 25 ms intervals for up to COMPRESS_TIMEOUT_MS (10 s) and returns as soon as the worker reports the compressed form, or returns the last- observed block on timeout so the caller's assertion still produces a meaningful failure. Convert the three call sites that used the fixed-sleep anti-pattern: - FedWorkerReadMatrixCompress.verifyRead (the actual CI flake) - FedWorkerMatrixCompress.verifySameOrAlsoCompressedAsLocalCompress (polls only when local compresses, so the "do not compress" parametrization stays fast) - FedWorkerMatrixMultiplyWorkload.verifySameOrAlsoCompressedAsLocalCompress Remove the now-unused FederatedTestUtils.wait helper so the anti-pattern is harder to reintroduce.
Fix intermittent ~26 minute hangs in the **.component.c**.** GitHub Actions job. The forks were finishing their test classes but failing to exit cleanly (leaked non-daemon threads from test executors), and Surefire 3.0.0 did not reliably enforce its fork shutdown timeout.
Bumps [actions/setup-node](https://github.com/actions/setup-node) from 6 to 7. - [Release notes](https://github.com/actions/setup-node/releases) - [Commits](actions/setup-node@v6...v7) --- updated-dependencies: - dependency-name: actions/setup-node dependency-version: '7' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Rewrites the Parquet frame reader to read columns via parquet's column API. Co-authored-by: Jakob-al28 <jakob28@mail.uni-paderborn.com>
Add three scenarios only an external engine can set up: a log that has rolled into a parquet checkpoint with a delete after it, DATE/TIMESTAMP/DECIMAL columns that have no SystemDS value type and must be rejected on the Delta type and a table SystemDS creates that Spark then appends to. Also stop countParquet from counting checkpoint files as data files.
This patch adds a min-max variance measure to Scuro's hyperparameter tuning functionality in order to avoid overfitting to the validation score. Additionally, it adds the missing parameter setters to multiple representations.
Assisted-by: AI
In this patch the node executor for Representation DAGs is improved and made mode efficient. Additionally, the code was cleaned up. Assisted-by: AI
Closes 2499. Signed-off-by: Grigorii Turchenko <turchenko.gregory@gmail.com>
Assisted-by: AI
Assisted-by: AI
Assisted-by: AI
…d test environments Closes apache#2587.
On newer windows version, the check for physical memory via wmic runs into a "cannot run program" error on hard crashes. This patch adds a more robust version that checks the modern way, old way, and in all cases prevents any crash because this physical memory size is just used for warnings that SystemDS runs in a JVM with too little memory.
Bumps [docker/login-action](https://github.com/docker/login-action) from 4.5.2 to 4.6.0. - [Release notes](https://github.com/docker/login-action/releases) - [Commits](docker/login-action@v4.5.2...v4.6.0) --- updated-dependencies: - dependency-name: docker/login-action dependency-version: 4.6.0 dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
This patch improves the representation DAGs, reduces duplicates and finds good initial parameters for window operators. Additionally, it collects more statistics of the optimization runs. Assisted-by: AI
This patch adds new representations for physiological data modalities, including new aggregators and context operations. Additionally, it refines existing operators with additional batch functionality and memory estimations. Assisted-by: AI
This patch adds improvements to multiple text, visual and audio representations. It includes a LazyLoading approach for visual modalities where the raw data is read when needed in the CustomDataloader. It adds two new representations: OpenFace and ImageBind. Assisted-by: AI
In this patch a new multimodal GA optimizer is added that implements various mutation operations. It reuses the worker pool and shared memory modules of the node executor. The GA optimizer uses the deap library. Assisted-by: AI
This patch adds functionality to apply representations on the test set only and perform inference on it. It measures the performance metrics as well as runtime for each step. Assisted-by: AI
This reverts commit 55e6bf6.
|
These aren't my changes in the conflict. Can I just accept all the incoming ones? |
This reverts commit 3962ff1.
Hi,
This is a homework pull-requests for the course AMLs at TUB.
This is also my first pull-request :)
I refactored
./scripts/algorithm/StepGLM.dmlinto a importable function.glm_fitwith adapter that uses the builtin solverm_gmlfrom./scripts/builtin/gml.dmlThe probit inverse link function is now approximated (AIC calculation):
Additionally:
./scripts/algorithms/TestBuiltinStepGLM.dmlperforms two checks for function output.Sources
[1]: Takeshi, Qualitative Response Models: A Survey (1981)