DL Random Search Exception - H2o 3.8.1.4
343 views
Skip to first unread message
Aarti Gaddale
Apr 8, 2016, 4:38:15 PM4/8/16
to H2O Open Source Scalable Machine Learning - h2ostream
Hi,
I get this exception while doing random search on DL, I see a similar issue on sparkling water (https://0xdata.atlassian.net/browse/PUBDEV-452)
Please help me understand what is going wrong here.
R Script:
hyper_params <- list(
activation=c("Rectifier","Tanh","Maxout","RectifierWithDropout","TanhWithDropout","MaxoutWithDropout"),
hidden=list(c(20,20),c(50,50),c(30,30,30),c(25,25,25,25)),
input_dropout_ratio=c(0,0.05),
l1=seq(0,1e-4,1e-6),
l2=seq(0,1e-4,1e-6)
)
hyper_params
## Stop once the top 5 models are within 1% of each other (i.e., the windowed average varies less than 1%)
search_criteria = list(strategy = "RandomDiscrete", max_runtime_secs = 360, max_models = 100, seed=1234567, stopping_rounds=5, stopping_tolerance=1e-2)
dl_random_grid <- h2o.grid(
algorithm="deeplearning",
grid_id = "dl_grid_random",
training_frame=train,
validation_frame=valid,
x=predictors,
y=response,
epochs=1,
stopping_metric="logloss",
stopping_tolerance=1e-2, ## stop when logloss does not improve by >=1% for 2 scoring events
stopping_rounds=2,
score_validation_samples=10000, ## downsample validation set for faster scoring
score_duty_cycle=0.025, ## don't score more than 2.5% of the wall time
max_w2=10, ## can help improve stability for Rectifier
hyper_params = hyper_params,
search_criteria = search_criteria,
ignored_columns=ignored_columns
)
Error 1
DistributedException from /127.0.0.1:54321
Parameters: [ 1.5E-5, 6.6E-5, 0.05, [50, 50], Rectifier ]
DistributedException from /127.0.0.1:54321, caused by java.lang.IllegalArgumentException: Operation not allowed on string vector.
at water.MRTask.getResult(MRTask.java:472) at water.MRTask.doAll(MRTask.java:389) at water.MRTask.doAll(MRTask.java:395) at hex.Model.predictScoreImpl(Model.java:852) at hex.deeplearning.DeepLearningModel.predictScoreImpl(DeepLearningModel.java:536) at hex.Model.score(Model.java:783) at hex.Model.score(Model.java:776) at hex.Model.score(Model.java:759) at hex.deeplearning.DeepLearningModel.doScoring(DeepLearningModel.java:395) at hex.deeplearning.DeepLearning$DeepLearningDriver.trainModel(DeepLearning.java:431) at hex.deeplearning.DeepLearning$DeepLearningDriver.buildModel(DeepLearning.java:302) at hex.deeplearning.DeepLearning$DeepLearningDriver.compute2(DeepLearning.java:203) at hex.ModelBuilder.trainModelNested(ModelBuilder.java:181) at hex.grid.GridSearch.startBuildModel(GridSearch.java:323) at hex.grid.GridSearch.buildModel(GridSearch.java:305) at hex.grid.GridSearch.gridSearch(GridSearch.java:209) at hex.grid.GridSearch.access$000(GridSearch.java:68) at hex.grid.GridSearch$1.compute2(GridSearch.java:134) at water.H2O$H2OCountedCompleter.compute(H2O.java:1181) at jsr166y.CountedCompleter.exec(CountedCompleter.java:468) at jsr166y.ForkJoinTask.doExec(ForkJoinTask.java:263) at jsr166y.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:974) at jsr166y.ForkJoinPool.runWorker(ForkJoinPool.java:1477) at jsr166y.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:104) Caused by: java.lang.IllegalArgumentException: Operation not allowed on string vector. at water.fvec.CStrChunk.atd_impl(CStrChunk.java:47) at water.fvec.Chunk.atd(Chunk.java:255) at hex.Model.score0(Model.java:963) at hex.Model$BigScore.map(Model.java:931) at water.MRTask.compute2(MRTask.java:637) at water.H2O$H2OCountedCompleter.compute1(H2O.java:1184) at hex.Model$BigScore$Icer.compute1(Model$BigScore$Icer.java) at water.H2O$H2OCountedCompleter.compute(H2O.java:1180) ... 5 more
Error 2
DistributedException from /127.0.0.1:54321
Parameters: [ 1.5E-5, 6.6E-5, 0.05, [50, 50], Rectifier ]
DistributedException from /127.0.0.1:54321, caused by java.lang.IllegalArgumentException: Operation not allowed on string vector.
at water.MRTask.getResult(MRTask.java:472) at water.MRTask.doAll(MRTask.java:389) at water.MRTask.doAll(MRTask.java:395) at hex.Model.predictScoreImpl(Model.java:852) at hex.deeplearning.DeepLearningModel.predictScoreImpl(DeepLearningModel.java:536) at hex.Model.score(Model.java:783) at hex.Model.score(Model.java:776) at hex.Model.score(Model.java:759) at hex.deeplearning.DeepLearningModel.doScoring(DeepLearningModel.java:395) at hex.deeplearning.DeepLearning$DeepLearningDriver.trainModel(DeepLearning.java:431) at hex.deeplearning.DeepLearning$DeepLearningDriver.buildModel(DeepLearning.java:302) at hex.deeplearning.DeepLearning$DeepLearningDriver.compute2(DeepLearning.java:203) at hex.ModelBuilder.trainModelNested(ModelBuilder.java:181) at hex.grid.GridSearch.startBuildModel(GridSearch.java:323) at hex.grid.GridSearch.buildModel(GridSearch.java:305) at hex.grid.GridSearch.gridSearch(GridSearch.java:209) at hex.grid.GridSearch.access$000(GridSearch.java:68) at hex.grid.GridSearch$1.compute2(GridSearch.java:134) at water.H2O$H2OCountedCompleter.compute(H2O.java:1181) at jsr166y.CountedCompleter.exec(CountedCompleter.java:468) at jsr166y.ForkJoinTask.doExec(ForkJoinTask.java:263) at jsr166y.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:974) at jsr166y.ForkJoinPool.runWorker(ForkJoinPool.java:1477) at jsr166y.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:104) Caused by: java.lang.IllegalArgumentException: Operation not allowed on string vector. at water.fvec.CStrChunk.atd_impl(CStrChunk.java:47) at water.fvec.Chunk.atd(Chunk.java:255) at hex.Model.score0(Model.java:963) at hex.Model$BigScore.map(Model.java:931) at water.MRTask.compute2(MRTask.java:637) at water.H2O$H2OCountedCompleter.compute1(H2O.java:1184) at hex.Model$BigScore$Icer.compute1(Model$BigScore$Icer.java) at water.H2O$H2OCountedCompleter.compute(H2O.java:1180) ... 5 more
Error 3
DistributedException from /127.0.0.1:54321
Parameters: [ 3.0E-6, 7.0E-5, 0.05, [20, 20], Rectifier ]
DistributedException from /127.0.0.1:54321, caused by java.lang.IllegalArgumentException: Operation not allowed on string vector.
at water.MRTask.getResult(MRTask.java:472) at water.MRTask.doAll(MRTask.java:389) at water.MRTask.doAll(MRTask.java:395) at hex.Model.predictScoreImpl(Model.java:852) at hex.deeplearning.DeepLearningModel.predictScoreImpl(DeepLearningModel.java:536) at hex.Model.score(Model.java:783) at hex.Model.score(Model.java:776) at hex.Model.score(Model.java:759) at hex.deeplearning.DeepLearningModel.doScoring(DeepLearningModel.java:395) at hex.deeplearning.DeepLearning$DeepLearningDriver.trainModel(DeepLearning.java:431) at hex.deeplearning.DeepLearning$DeepLearningDriver.buildModel(DeepLearning.java:302) at hex.deeplearning.DeepLearning$DeepLearningDriver.compute2(DeepLearning.java:203) at hex.ModelBuilder.trainModelNested(ModelBuilder.java:181) at hex.grid.GridSearch.startBuildModel(GridSearch.java:323) at hex.grid.GridSearch.buildModel(GridSearch.java:305) at hex.grid.GridSearch.gridSearch(GridSearch.java:209) at hex.grid.GridSearch.access$000(GridSearch.java:68) at hex.grid.GridSearch$1.compute2(GridSearch.java:134) at water.H2O$H2OCountedCompleter.compute(H2O.java:1181) at jsr166y.CountedCompleter.exec(CountedCompleter.java:468) at jsr166y.ForkJoinTask.doExec(ForkJoinTask.java:263) at jsr166y.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:974) at jsr166y.ForkJoinPool.runWorker(ForkJoinPool.java:1477) at jsr166y.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:104) Caused by: java.lang.IllegalArgumentException: Operation not allowed on string vector. at water.fvec.CStrChunk.atd_impl(CStrChunk.java:47) at water.fvec.Chunk.atd(Chunk.java:255) at hex.Model.score0(Model.java:963) at hex.Model$BigScore.map(Model.java:931) at water.MRTask.compute2(MRTask.java:637) at water.H2O$H2OCountedCompleter.compute1(H2O.java:1184) at hex.Model$BigScore$Icer.compute1(Model$BigScore$Icer.java) at water.H2O$H2OCountedCompleter.compute(H2O.java:1180) ... 5 more
Error 4
DistributedException from /127.0.0.1:54321
Parameters: [ 9.6E-5, 6.9E-5, 0.0, [25, 25, 25, 25], MaxoutWithDropout ]
DistributedException from /127.0.0.1:54321, caused by java.lang.IllegalArgumentException: Operation not allowed on string vector.
at water.MRTask.getResult(MRTask.java:472) at water.MRTask.doAll(MRTask.java:389) at water.MRTask.doAll(MRTask.java:395) at hex.Model.predictScoreImpl(Model.java:852) at hex.deeplearning.DeepLearningModel.predictScoreImpl(DeepLearningModel.java:536) at hex.Model.score(Model.java:783) at hex.Model.score(Model.java:776) at hex.Model.score(Model.java:759) at hex.deeplearning.DeepLearningModel.doScoring(DeepLearningModel.java:395) at hex.deeplearning.DeepLearning$DeepLearningDriver.trainModel(DeepLearning.java:431) at hex.deeplearning.DeepLearning$DeepLearningDriver.buildModel(DeepLearning.java:302) at hex.deeplearning.DeepLearning$DeepLearningDriver.compute2(DeepLearning.java:203) at hex.ModelBuilder.trainModelNested(ModelBuilder.java:181) at hex.grid.GridSearch.startBuildModel(GridSearch.java:323) at hex.grid.GridSearch.buildModel(GridSearch.java:305) at hex.grid.GridSearch.gridSearch(GridSearch.java:209) at hex.grid.GridSearch.access$000(GridSearch.java:68) at hex.grid.GridSearch$1.compute2(GridSearch.java:134) at water.H2O$H2OCountedCompleter.compute(H2O.java:1181) at jsr166y.CountedCompleter.exec(CountedCompleter.java:468) at jsr166y.ForkJoinTask.doExec(ForkJoinTask.java:263) at jsr166y.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:974) at jsr166y.ForkJoinPool.runWorker(ForkJoinPool.java:1477) at jsr166y.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:104) Caused by: java.lang.IllegalArgumentException: Operation not allowed on string vector. at water.fvec.CStrChunk.atd_impl(CStrChunk.java:47) at water.fvec.Chunk.atd(Chunk.java:255) at hex.Model.score0(Model.java:963) at hex.Model$BigScore.map(Model.java:931) at water.MRTask.compute2(MRTask.java:637) at water.H2O$H2OCountedCompleter.compute1(H2O.java:1184) at hex.Model$BigScore$Icer.compute1(Model$BigScore$Icer.java) at water.H2O$H2OCountedCompleter.compute(H2O.java:1180) ... 5 more
Michal Malohlava
Apr 11, 2016, 8:33:19 PM4/11/16
to h2os...@googlegroups.com
Please make sure you converted all your
string-features to categorical features.
Thank you!
Michal
Thank you!
Michal
--
You received this message because you are subscribed to the Google Groups "H2O Open Source Scalable Machine Learning - h2ostream" group.
To unsubscribe from this group and stop receiving emails from it, send an email to h2ostream+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
nachike...@linaro.org
May 27, 2016, 5:34:06 PM5/27/16
to H2O Open Source Scalable Machine Learning - h2ostream, mic...@h2oai.com
Hi Michal,
Can you please elaborate on how to do this conversion? I am facing this exception on Sparkling Water 1.6.3.
I followed the steps in the Craigslist demo video (https://www.youtube.com/watch?v=HVairux8LUM) and am getting this same exception when I build the model:
DistributedException from localhost/127.0.0.1:54321, caused by java.lang.IllegalArgumentException: Operation not allowed on string vector.
I am working on an ARM64 system with Linaro OpenJDK 1.8.
Please let me know.
Thanks,
Nachiket
Can you please elaborate on how to do this conversion? I am facing this exception on Sparkling Water 1.6.3.
I followed the steps in the Craigslist demo video (https://www.youtube.com/watch?v=HVairux8LUM) and am getting this same exception when I build the model:
DistributedException from localhost/127.0.0.1:54321, caused by java.lang.IllegalArgumentException: Operation not allowed on string vector.
I am working on an ARM64 system with Linaro OpenJDK 1.8.
Please let me know.
Thanks,
Nachiket
jack...@gmail.com
Sep 8, 2016, 9:53:43 AM9/8/16
to H2O Open Source Scalable Machine Learning - h2ostream, mic...@h2oai.com, nachike...@linaro.org
суббота, 28 мая 2016 г., 0:34:06 UTC+3 пользователь nachike...@linaro.org написал:
Hi, Aarti.
You should change column type of target to enum. It's string by default. U can do it by Get frames -> inspect your frame - click on "change to enum" link in the row with target. Or u can execute cs "changeColumnType frame: "frame_rdd_52", column: "target", type: 'enum'" just change frame_rdd_52 to your frame name.
You should change column type of target to enum. It's string by default. U can do it by Get frames -> inspect your frame - click on "change to enum" link in the row with target. Or u can execute cs "changeColumnType frame: "frame_rdd_52", column: "target", type: 'enum'" just change frame_rdd_52 to your frame name.
Reply all
Reply to author
Forward
0 new messages