revert back api (intel-analytics#2943)

dding3 · Oct 24, 2019 · 1c7f742 · 1c7f742
1 parent dff06ce
commit 1c7f742
Show file tree

Hide file tree

Showing 3 changed files with 15 additions and 34 deletions.
diff --git a/...a/dllib/src/main/scala/com/intel/analytics/bigdl/dllib/nn/abstractnn/AbstractModule.scala b/...a/dllib/src/main/scala/com/intel/analytics/bigdl/dllib/nn/abstractnn/AbstractModule.scala
@@ -647,7 +647,7 @@ abstract class AbstractModule[A <: Activity: ClassTag, B <: Activity: ClassTag,
    *                  if -1, default is 4 * partitionNumber of dataset
    */
 
-  final def predictClass(dataset: RDD[Sample[T]], batchSize: Int = -1): RDD[Sample[T]] = {
+  final def predictClass(dataset: RDD[Sample[T]], batchSize: Int = -1): RDD[Int] = {
     Predictor(this).predictClass(dataset, batchSize)
   }
 

diff --git a/scala/dllib/src/main/scala/com/intel/analytics/bigdl/dllib/optim/Predictor.scala b/scala/dllib/src/main/scala/com/intel/analytics/bigdl/dllib/optim/Predictor.scala
@@ -180,34 +180,15 @@ object Predictor {
 
   def predictClass[T: ClassTag](dataSet: RDD[Sample[T]], batchSize: Int = -1, model: Module[T],
     batchPerPartition: Int, featurePaddingParam: Option[PaddingParam[T]])(
-    implicit ev: TensorNumeric[T]): RDD[Sample[T]] = {
-    val shareBuffer = false
-    val modelBroad = ModelBroadcast[T]().broadcast(dataSet.sparkContext,
-      ConversionUtils.convert(model.evaluate()))
-    val partitionNum = dataSet.partitions.length
-    val totalBatch = if (batchSize > 0) {
-      require(batchSize % partitionNum == 0, s"Predictor.predict: total batch size $batchSize " +
-        s"should be divided by partitionNum ${partitionNum}")
-      batchSize
-    } else {
-      batchPerPartition * partitionNum
-    }
-    val rdd = ConversionUtils.coalesce(dataSet)
-    val realPartitionLength = rdd.partitions.length
-    val otherBroad = rdd.sparkContext.broadcast(SampleToMiniBatch(
-      batchSize = totalBatch,
-      partitionNum = Some(realPartitionLength),
-      featurePaddingParam = featurePaddingParam))
-    val localBatchPerPartition = totalBatch / realPartitionLength
-    rdd.mapPartitions { partition =>
-      val localModel = modelBroad.value()
-      val localTransformer = otherBroad.value.cloneTransformer()
-      partition.grouped(localBatchPerPartition).flatMap(samples => {
-        val batchOut = predictSamples(localModel, samples, localTransformer, shareBuffer)
-        samples.toIterator.zip(batchOut).foreach(tuple => {
-          Sample(tuple._1.feature(), tuple._2.toTensor)
-        })
-        samples
+    implicit ev: TensorNumeric[T]): RDD[Int] = {
+    val result = Predictor.predict(dataSet, batchSize, true, model,
+      batchPerPartition, featurePaddingParam)
+    result.mapPartitions { partition =>
+      partition.map(output => {
+        val _output = output.toTensor[T]
+        require(_output.dim() == 1, s"Predictor.predictClass:" +
+          s"Only support one sample has one label, but got ${_output.dim()} label")
+        ev.toType[Int](_output.max(1)._2.valueAt(1))
       })
     }
   }
@@ -233,7 +214,7 @@ class Predictor[T: ClassTag] private[optim](
    batchPerPartition: Int = 4)
   (implicit ev: TensorNumeric[T]) extends Serializable {
 
-  def predictClass(dataSet: RDD[Sample[T]], batchSize: Int = -1): RDD[Sample[T]] = {
+  def predictClass(dataSet: RDD[Sample[T]], batchSize: Int = -1): RDD[Int] = {
     Predictor.predictClass(dataSet, batchSize, model, batchPerPartition, featurePaddingParam)
   }
 

diff --git a/...a/dllib/src/main/scala/com/intel/analytics/bigdl/dllib/utils/python/api/PythonBigDL.scala b/...a/dllib/src/main/scala/com/intel/analytics/bigdl/dllib/utils/python/api/PythonBigDL.scala
@@ -2050,10 +2050,10 @@ class PythonBigDL[T: ClassTag](implicit ev: TensorNumeric[T]) extends Serializab
   }
 
   def modelPredictClass(model: AbstractModule[Activity, Activity, T],
-                      dataRdd: JavaRDD[Sample]): JavaRDD[Sample] = {
-    val sampleRDD = toJSample(dataRdd)
-    val pySampleRDD = model.predictClass(sampleRDD).map(toPySample(_))
-    new JavaRDD[Sample](pySampleRDD)
+                        dataRdd: JavaRDD[Sample]): JavaRDD[Int] = {
+    val sampleRdd = toJSample(dataRdd)
+    val tensorRDD = model.predictClass(sampleRdd)
+    new JavaRDD[Int](tensorRDD)
   }
 
   def modelForward(model: AbstractModule[Activity, Activity, T],