Repository navigation
Closed
Milestone
Description
System Information (please complete the following information):
- OS & Version: Windows 11 [Version 10.0.22621.675]
- ML.NET Version: ML.NT v2.0.0-preview.22551.1
- .NET Version: .NET 6.0.9
Describe the bug
LightGBM trials fail with "bad allocation" error message and they cause OutOfMemoryException even if SetMaximumMemoryUsageInMegaByte() is used.
To Reproduce
Steps to reproduce the behavior:
- Create a pipeline that includes LightGBM, for example:
var pipeline =
mlContext.Auto().Featurizer(trainTestData.TrainSet, numericColumns: new[] { "Features" })
.Append(mlContext.Auto().Regression(useFastTree: false, useLbfgs: true, useSdca: true, useFastForest: false, useLgbm: true));
- Create an ML.NET experiment and set its memory limit. I set it to 20 GB, because I usually have ~24 GB available when I start the training.
var experiment = mlContext.Auto().CreateExperiment();
experiment
.SetPipeline(pipeline)
.SetTrainingTimeInSeconds(trainingTimeInSeconds)
.SetRegressionMetric(RegressionMetric.RSquared, labelColumn: "Label")
.SetDataset(trainTestData.TrainSet, trainTestData.TestSet)
.SetMonitor(monitor)
.SetMaximumMemoryUsageInMegaByte(20 * 1024);
- Run the experiment for an hour
- The experiment will end with
OutOfMemoryException
Expected behavior
I would expect the AutoML stops the trial if it exceeds the memory limit and/or there is an exception during its run. It should not about the whole experiment.
Screenshots, Code, Sample Projects
Here are the logs of three runs I had today:
15:23:11.3 info: Bio Balance Detector Body Monitor v0.9.1.3
15:23:11.3 info: (The current UTC time is 2022-11-02 14:23:11)
15:23:11.6 info: Now listening on: https://0.0.0.0:7061
15:23:11.6 info: Now listening on: http://localhost:5061
15:23:11.6 info: Microsoft.Hosting.Lifetime[0] Application started. Press Ctrl+C to shut down.
15:23:11.6 info: Microsoft.Hosting.Lifetime[0] Hosting environment: Development
15:23:11.6 info: Microsoft.Hosting.Lifetime[0] Content root path: C:\Work\BioBalanceDetector\Software\Source\BBDProto08\BBD.BodyMonitor\BBD.BodyMonitor.API\
15:23:38.1 info: Starting model training for 3600 seconds, using 'BBD_20221102_2022-11-02_MLP14_0p25Hz-6250Hz_IsSession.SegmentedData.Sleep.Level_1098rows.csv' as data source with the 'MLP14' profile.
15:23:51.5 info: Completed Trial # 0 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,15 - Duration: 13 seconds
15:23:51.5 info: New Best Trial # 0 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,15
15:24:01.2 info: Completed Trial # 1 - Pipeline: ReplaceMissingValues=>Concatenate=>LightGbmRegression - Metric: 0,05 - Duration: 10 seconds
15:24:11.1 info: Completed Trial # 2 - Pipeline: ReplaceMissingValues=>Concatenate=>FastTreeRegression - Metric: -2,20 - Duration: 10 seconds
15:50:15.5 info: Completed Trial # 3 - Pipeline: ReplaceMissingValues=>Concatenate=>SdcaRegression - Metric: 0,03 - Duration: 1564 seconds
15:50:32.2 info: Completed Trial # 4 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,41 - Duration: 17 seconds
15:50:32.2 info: New Best Trial # 4 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,41
15:50:44.3 info: Completed Trial # 5 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,16 - Duration: 12 seconds
15:50:50.0 info: Completed Trial # 6 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,40 - Duration: 6 seconds
15:51:15.3 info: Completed Trial # 7 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,43 - Duration: 25 seconds
15:51:15.3 info: New Best Trial # 7 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,43
15:51:43.3 info: Completed Trial # 8 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,43 - Duration: 28 seconds
15:51:43.3 info: New Best Trial # 8 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,43
15:51:54.4 info: Completed Trial # 9 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,14 - Duration: 11 seconds
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
17:49:52.6 info: BBD.BodyMonitor.API[0] Bio Balance Detector Body Monitor v0.9.1.3
17:49:52.6 info: BBD.BodyMonitor.API[0] (The current UTC time is 2022-11-02 16:49:52)
17:49:52.9 info: Microsoft.Hosting.Lifetime[14] Now listening on: https://0.0.0.0:7061
17:49:52.9 info: Microsoft.Hosting.Lifetime[14] Now listening on: http://localhost:5061
17:49:52.9 info: Microsoft.Hosting.Lifetime[0] Application started. Press Ctrl+C to shut down.
17:49:52.9 info: Microsoft.Hosting.Lifetime[0] Hosting environment: Development
17:49:52.9 info: Microsoft.Hosting.Lifetime[0] Content root path: C:\Work\BioBalanceDetector\Software\Source\BBDProto08\BBD.BodyMonitor\BBD.BodyMonitor.API\
17:50:07.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Starting model training for 3600 seconds, using 'BBD_20221102_2022-11-02_MLP14_0p25Hz-6250Hz_IsSession.SegmentedData.Sleep.Level_1098rows.csv' as data source with the 'MLP14' profile.
17:50:08.0 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 0 - Pipeline: ReplaceMissingValues=>Concatenate=>FastTreeRegression
17:50:20.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 0 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,15 - Duration: 13 seconds
17:50:20.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] New Best Trial # 0 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,15
17:50:20.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 1 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
17:50:30.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 1 - Pipeline: ReplaceMissingValues=>Concatenate=>FastTreeRegression - Metric: -2,20 - Duration: 10 seconds
17:50:30.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 2 - Pipeline: ReplaceMissingValues=>Concatenate=>FastTreeRegression
17:50:40.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 2 - Pipeline: ReplaceMissingValues=>Concatenate=>LightGbmRegression - Metric: 0,05 - Duration: 10 seconds
17:50:40.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 3 - Pipeline: ReplaceMissingValues=>Concatenate=>LightGbmRegression
18:16:49.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 3 - Pipeline: ReplaceMissingValues=>Concatenate=>SdcaRegression - Metric: 0,03 - Duration: 1569 seconds
18:16:49.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 4 - Pipeline: ReplaceMissingValues=>Concatenate=>SdcaRegression
18:17:05.9 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 4 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,41 - Duration: 17 seconds
18:17:05.9 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] New Best Trial # 4 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,41
18:17:05.9 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 5 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:17:17.9 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 5 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,21 - Duration: 12 seconds
18:17:17.9 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 6 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
18:17:22.7 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 6 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,41 - Duration: 5 seconds
18:17:22.7 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 7 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:17:34.3 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 7 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,23 - Duration: 12 seconds
18:17:34.3 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 8 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
18:17:50.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 8 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,36 - Duration: 16 seconds
18:17:50.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 9 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:19:58.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 9 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,28 - Duration: 128 seconds
18:19:58.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 10 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
18:21:04.9 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 10 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,26 - Duration: 67 seconds
18:21:04.9 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 11 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
18:21:18.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 11 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,22 - Duration: 14 seconds
18:21:18.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 12 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:21:23.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 12 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,36 - Duration: 5 seconds
18:21:23.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 13 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:21:35.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 13 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,40 - Duration: 12 seconds
18:21:35.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 14 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:21:58.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 14 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,43 - Duration: 23 seconds
18:21:58.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] New Best Trial # 14 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,43
18:21:58.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 15 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:22:14.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 15 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,27 - Duration: 16 seconds
18:22:14.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 16 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:25:06.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 16 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,28 - Duration: 172 seconds
18:25:06.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 17 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
18:25:11.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 17 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,39 - Duration: 5 seconds
18:25:11.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 18 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:25:16.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 18 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,38 - Duration: 5 seconds
18:25:16.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 19 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:25:22.0 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 19 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,03 - Duration: 6 seconds
18:25:22.0 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 20 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:25:32.3 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 20 - Pipeline: ReplaceMissingValues=>Concatenate=>LightGbmRegression - Metric: 0,02 - Duration: 10 seconds
18:25:32.3 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 21 - Pipeline: ReplaceMissingValues=>Concatenate=>LightGbmRegression
18:25:37.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 21 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,37 - Duration: 5 seconds
18:25:37.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 22 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
18:26:03.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 22 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,29 - Duration: 26 seconds
18:26:03.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 23 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
18:26:07.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 23 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,04 - Duration: 4 seconds
18:26:07.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 24 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
18:53:28.1 info: BBD.BodyMonitor.API[0] Bio Balance Detector Body Monitor v0.9.1.3
18:53:28.1 info: BBD.BodyMonitor.API[0] (The current UTC time is 2022-11-02 17:53:28)
18:53:28.6 info: Microsoft.Hosting.Lifetime[14] Now listening on: https://0.0.0.0:7061
18:53:28.6 info: Microsoft.Hosting.Lifetime[14] Now listening on: http://localhost:5061
18:53:28.6 info: Microsoft.Hosting.Lifetime[0] Application started. Press Ctrl+C to shut down.
18:53:28.6 info: Microsoft.Hosting.Lifetime[0] Hosting environment: Development
18:53:28.6 info: Microsoft.Hosting.Lifetime[0] Content root path: C:\Work\BioBalanceDetector\Software\Source\BBDProto08\BBD.BodyMonitor\BBD.BodyMonitor.API\
18:53:44.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Starting model training for 3600 seconds, using 'BBD_20221102_2022-11-02_MLP14_0p25Hz-6250Hz_IsSession.SegmentedData.Sleep.Level_1098rows.csv' as data source with the 'MLP14' profile.
18:53:44.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 0 - Pipeline: ReplaceMissingValues=>Concatenate=>FastTreeRegression
18:53:58.7 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 0 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,15 - Duration: 14 seconds
18:53:58.7 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] New Best Trial # 0 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,15
18:53:58.7 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 1 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
18:54:08.8 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 1 - Pipeline: ReplaceMissingValues=>Concatenate=>LightGbmRegression - Metric: 0,05 - Duration: 10 seconds
18:54:08.8 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 2 - Pipeline: ReplaceMissingValues=>Concatenate=>LightGbmRegression
19:19:50.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 2 - Pipeline: ReplaceMissingValues=>Concatenate=>SdcaRegression - Metric: 0,03 - Duration: 1542 seconds
19:19:50.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 3 - Pipeline: ReplaceMissingValues=>Concatenate=>SdcaRegression
19:20:00.8 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 3 - Pipeline: ReplaceMissingValues=>Concatenate=>FastTreeRegression - Metric: -2,20 - Duration: 10 seconds
19:20:00.8 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 4 - Pipeline: ReplaceMissingValues=>Concatenate=>FastTreeRegression
19:20:17.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 4 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,41 - Duration: 17 seconds
19:20:17.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] New Best Trial # 4 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,41
19:20:17.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 5 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:20:32.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 5 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,25 - Duration: 15 seconds
19:20:32.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 6 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
19:21:19.0 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 6 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,26 - Duration: 47 seconds
19:21:19.0 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 7 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
19:21:41.8 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 7 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,39 - Duration: 23 seconds
19:21:41.8 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 8 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:21:46.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 8 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,37 - Duration: 5 seconds
19:21:46.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 9 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:22:04.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 9 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,23 - Duration: 18 seconds
19:22:04.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 10 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
19:22:08.9 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 10 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,41 - Duration: 5 seconds
19:22:08.9 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 11 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:22:26.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 11 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,31 - Duration: 18 seconds
19:22:26.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 12 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:22:31.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 12 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,36 - Duration: 5 seconds
19:22:31.2 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 13 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:22:58.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 13 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,35 - Duration: 27 seconds
19:22:58.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 14 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:23:04.3 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 14 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,40 - Duration: 6 seconds
19:23:04.3 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 15 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:23:20.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 15 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,24 - Duration: 16 seconds
19:23:20.4 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 16 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:23:37.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 16 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,41 - Duration: 17 seconds
19:23:37.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 17 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:23:46.7 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 17 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,39 - Duration: 9 seconds
19:23:46.7 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 18 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:24:14.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 18 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,43 - Duration: 28 seconds
19:24:14.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] New Best Trial # 18 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,43
19:24:14.6 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 19 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:24:46.8 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 19 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,43 - Duration: 32 seconds
19:24:46.8 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 20 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:24:57.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 20 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,41 - Duration: 10 seconds
19:24:57.1 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 21 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:25:14.3 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 21 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression - Metric: 0,29 - Duration: 17 seconds
19:25:14.3 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 22 - Pipeline: ReplaceMissingValues=>Concatenate=>LbfgsPoissonRegressionRegression
19:26:55.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 22 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression - Metric: 0,27 - Duration: 101 seconds
19:26:55.5 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 23 - Pipeline: ReplaceMissingValues=>Concatenate=>FastForestRegression
19:27:05.0 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Completed Trial # 23 - Pipeline: ReplaceMissingValues=>Concatenate=>LightGbmRegression - Metric: 0,13 - Duration: 10 seconds
19:27:05.0 info: BBD.BodyMonitor.API.Controllers.MachineLearningController[0] Running Trial # 24 - Pipeline: ReplaceMissingValues=>Concatenate=>LightGbmRegression
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
[LightGBM] [Warning] bad allocation
Additional context
My project is open-source so both the source and the data are available. Let me know if you need them for testing.
Metadata
Metadata
Assignees
Labels
No labels
Activity
michaelgsharp commentedon Nov 28, 2022
@LittleLittleCloud familiar with this?
LittleLittleCloud commentedon Nov 29, 2022
The warning message should be from lightGBM unmanaged code while
SetMaximumMemoryUsageInMegaByteonly restricts memory usage in managed code. So ifOutOfMemoryhappens in unmanaged codeAutoML.Netwon't have the ability to cancel that trial.@andrasfuchs can you share with us the size of your machine.
andrasfuchs commentedon Dec 2, 2022
@LittleLittleCloud The critical resource in this regard is the size of the virtual memory in my system. I have 32 GB RAM and ~43 GB of virtual memory. I start to get this
bad allocationwarning if there is no more free virtual memory left.I don't know how virtual memory works in other OSs, but I used this Windows-specific method in my
IMonitorto monitor it during my trials:(Note: I use this check in my modified ML.NET code, where the
IMonitorhas aReportTrialResourceUsage(TrialSettings settings)method that gets called repeatedly during trial runs to allow continuous monitoring.)luisquintanilla commentedon Dec 7, 2023
Closing this issue since it should've been addressed in #6520.