===== MAIN: learn based on training data =====
=== START program1: ./run learn ../dataset3/train
=== START program2: ./run learn ../program1/data
main() {
Reading examples from ../program1/data {
193657 examples, 520 features, 2 labels
} [25s, cum. 25s]
Iteration 0 {
numMistakes = 72606/193657 = 0.375
} [3.6s, cum. 28s]
Iteration 1 {
numMistakes = 72162/193657 = 0.373
} [6.2s, cum. 35s]
Iteration 2 {
numMistakes = 71247/193657 = 0.368
} [3.6s, cum. 38s]
Iteration 3 {
numMistakes = 70773/193657 = 0.365
} [3.7s, cum. 42s]
Iteration 4 {
numMistakes = 70055/193657 = 0.362
} [3.7s, cum. 46s]
Writing parameters to params
} [46s]
=== END program2: ./run learn ../program1/data --- OK [47s]
=== END program1: ./run learn ../dataset3/train --- OK [73s]
===== MAIN: predict/evaluate on train data =====
=== START program4: ./run stripLabels ../dataset3/train ../program0/evalTrain.in
=== END program4: ./run stripLabels ../dataset3/train ../program0/evalTrain.in --- OK [5s]
=== START program1: ./run predict ../program0/evalTrain.in ../program0/evalTrain.out
=== START program2: ./run predict ../program0/evalTrain.in ../program0/evalTrain.out.multiclass-output
main() {
Reading parameters from params
Reading examples from ../program0/evalTrain.in {
193657 examples, 520 features, 2 labels
} [24s, cum. 24s]
Predicting [3.5s, cum. 28s]
} [28s]
=== END program2: ./run predict ../program0/evalTrain.in ../program0/evalTrain.out.multiclass-output --- OK [29s]
=== END program1: ./run predict ../program0/evalTrain.in ../program0/evalTrain.out --- OK [36s]
=== START program5: ./run evaluate ../dataset3/train ../program0/evalTrain.out
=== END program5: ./run evaluate ../dataset3/train ../program0/evalTrain.out --- OK [20s]
===== MAIN: predict/evaluate on test data =====
=== START program4: ./run stripLabels ../dataset3/test ../program0/evalTest.in
=== END program4: ./run stripLabels ../dataset3/test ../program0/evalTest.in --- OK [3s]
=== START program1: ./run predict ../program0/evalTest.in ../program0/evalTest.out
=== START program2: ./run predict ../program0/evalTest.in ../program0/evalTest.out.multiclass-output
main() {
Reading parameters from params
Reading examples from ../program0/evalTest.in {
82996 examples, 520 features, 2 labels
} [11s, cum. 11s]
Predicting
} [12s]
=== END program2: ./run predict ../program0/evalTest.in ../program0/evalTest.out.multiclass-output --- OK [13s]
=== END program1: ./run predict ../program0/evalTest.in ../program0/evalTest.out --- OK [16s]
=== START program5: ./run evaluate ../dataset3/test ../program0/evalTest.out
=== END program5: ./run evaluate ../dataset3/test ../program0/evalTest.out --- OK [8s]
real 2m54.373s
user 2m27.253s
sys 0m12.949s
Run specification
supervised-learning: Main entry for supervised learning for training and testing a program on a dataset.
(learner:Program) binary-to-multi: Allows multiclass classification to be run on binary classification datasets (trivial reduction).
(multiclassLearner:Program[MulticlassClassification]) sgd-logistic-stepsizeAdaGrad-iter5: Stochastic gradient descent (loss=logistic, stepSize = AdaGrad, take 5 passes over training data)
(dataset:Dataset) dominionstats - large: 270k intermediate game states from the online Dominion game server at dominion.isotropic.org. You need to predict the winner.
A small version of this data set is here: http://mlcomp.org/datasets/871
(stripper:Program[Strip]) binary-utils: Validates and inspects a dataset in BinaryClassification format.
(evaluator:Program[Evaluate]) classification-evaluator: Evaluates predictions of classification datasets (discrete outputs).
When you generate a run, you can set a time limit for the run (no more than 24 hours). After that point, we will terminate the program.
Your program can use 1.5GB of memory. More information here.
Go to the page for the run and look at the log file for signs of the responsible error.
You can also download the run and run it locally on your machine (a README file should
be included in the download which provides more information).
We said that a run was simply a program/dataset pair, but that's not the full story.
A run actually includes other helper programs such as the evaluation program and
various programs for reductions (e.g., one-versus-all, hyperparameter tuning).
More formally, a run is a given by a run specification,
which can be found on the page for any run.
A run specification is a tree where each internal node represents a program
and its children represents the arguments to be passed into its constructor.
For example, the one-versus-all program takes your binary classification program
as a constructor argument and behaves like a multiclass classification program.
Must be logged in to post comments.