Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

change the API name since we support not only HDFS and local file system

Browse Source
This commit is contained in:
CodingCat 2016-03-11 10:00:37 -05:00
parent 8e3ce908fe
commit 43d7a85bc9
4 changed files with 6 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
jvm-packages/xgboost4j-example/src/main/scala/ml/dmlc/xgboost4j/scala/example/spark/DistTrainWithSpark.scala
View File

@ -40,6 +40,6 @@ object DistTrainWithSpark {
"objective" -> "binary:logistic").toMap "objective" -> "binary:logistic").toMap
val xgboostModel = XGBoost.train(trainRDD, paramMap, numRound) val xgboostModel = XGBoost.train(trainRDD, paramMap, numRound)
// save model to HDFS path // save model to HDFS path
xgboostModel.saveModelToHadoop(outputModelPath) xgboostModel.saveModelAsHadoopFile(outputModelPath)
} }
} }

3
jvm-packages/xgboost4j-spark/src/main/scala/ml/dmlc/xgboost4j/scala/spark/XGBoost.scala
View File

@ -128,7 +128,8 @@ object XGBoost extends Serializable {
* @param modelPath The path of the file representing the model * @param modelPath The path of the file representing the model
* @return The loaded model * @return The loaded model
*/ */
def loadModelFromHadoop(modelPath: String)(implicit sparkContext: SparkContext): XGBoostModel = { def loadModelFromHadoopFile(modelPath: String)(implicit sparkContext: SparkContext):
XGBoostModel = {
val path = new Path(modelPath) val path = new Path(modelPath)
val dataInStream = path.getFileSystem(sparkContext.hadoopConfiguration).open(path) val dataInStream = path.getFileSystem(sparkContext.hadoopConfiguration).open(path)
val xgBoostModel = new XGBoostModel(SXGBoost.loadModel(dataInStream)) val xgBoostModel = new XGBoostModel(SXGBoost.loadModel(dataInStream))

2
jvm-packages/xgboost4j-spark/src/main/scala/ml/dmlc/xgboost4j/scala/spark/XGBoostModel.scala
View File

@ -49,7 +49,7 @@ class XGBoostModel(booster: Booster)(implicit val sc: SparkContext) extends Seri
* *
* @param modelPath The model path as in Hadoop path. * @param modelPath The model path as in Hadoop path.
*/ */
def saveModelToHadoop(modelPath: String): Unit = { def saveModelToHadoopFile(modelPath: String): Unit = {
val path = new Path(modelPath) val path = new Path(modelPath)
val outputStream = path.getFileSystem(sc.hadoopConfiguration).create(path) val outputStream = path.getFileSystem(sc.hadoopConfiguration).create(path)
booster.saveModel(outputStream) booster.saveModel(outputStream)

4
jvm-packages/xgboost4j-spark/src/test/scala/ml/dmlc/xgboost4j/scala/spark/XGBoostSuite.scala
View File

@ -150,8 +150,8 @@ class XGBoostSuite extends FunSuite with BeforeAndAfter {
"objective" -> "binary:logistic").toMap "objective" -> "binary:logistic").toMap
val xgBoostModel = XGBoost.train(trainingRDD, paramMap, 5) val xgBoostModel = XGBoost.train(trainingRDD, paramMap, 5)
assert(eval.eval(xgBoostModel.predict(testSetDMatrix), testSetDMatrix) < 0.1) assert(eval.eval(xgBoostModel.predict(testSetDMatrix), testSetDMatrix) < 0.1)
xgBoostModel.saveModelToHadoop(tempFile.toFile.getAbsolutePath) xgBoostModel.saveModelAsHadoopFile(tempFile.toFile.getAbsolutePath)
val loadedXGBooostModel = XGBoost.loadModelFromHadoop(tempFile.toFile.getAbsolutePath) val loadedXGBooostModel = XGBoost.loadModelFromHadoopFile(tempFile.toFile.getAbsolutePath)
val predicts = loadedXGBooostModel.predict(testSetDMatrix) val predicts = loadedXGBooostModel.predict(testSetDMatrix)
assert(eval.eval(predicts, testSetDMatrix) < 0.1) assert(eval.eval(predicts, testSetDMatrix) < 0.1)
} }