-
Notifications
You must be signed in to change notification settings - Fork 729
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
support multi input models for nnframes (#1553)
* support multi input for nnframes * update ut * add doc and unit test * doc update * scala style
- Loading branch information
Showing
11 changed files
with
406 additions
and
23 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
38 changes: 38 additions & 0 deletions
38
zoo/src/main/scala/com/intel/analytics/zoo/feature/common/MultiTensorsToSample.scala
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,38 @@ | ||
| /* | ||
| * Copyright 2018 Analytics Zoo Authors. | ||
| * | ||
| * Licensed under the Apache License, Version 2.0 (the "License"); | ||
| * you may not use this file except in compliance with the License. | ||
| * You may obtain a copy of the License at | ||
| * | ||
| * http://www.apache.org/licenses/LICENSE-2.0 | ||
| * | ||
| * Unless required by applicable law or agreed to in writing, software | ||
| * distributed under the License is distributed on an "AS IS" BASIS, | ||
| * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
| * See the License for the specific language governing permissions and | ||
| * limitations under the License. | ||
| */ | ||
| package com.intel.analytics.zoo.feature.common | ||
|
|
||
| import com.intel.analytics.bigdl.dataset.Sample | ||
| import com.intel.analytics.bigdl.tensor.Tensor | ||
| import com.intel.analytics.bigdl.tensor.TensorNumericMath.TensorNumeric | ||
|
|
||
| import scala.reflect.ClassTag | ||
|
|
||
| /** | ||
| * a Preprocessing that converts multiple tensors to Sample. | ||
| */ | ||
| class MultiTensorsToSample[T: ClassTag]()(implicit ev: TensorNumeric[T]) | ||
| extends Preprocessing[Array[Tensor[T]], Sample[T]] { | ||
|
|
||
| override def apply(prev: Iterator[Array[Tensor[T]]]): Iterator[Sample[T]] = { | ||
| prev.map(Sample(_)) | ||
| } | ||
| } | ||
|
|
||
| object MultiTensorsToSample { | ||
| def apply[F, T: ClassTag]()(implicit ev: TensorNumeric[T]): MultiTensorsToSample[T] = | ||
| new MultiTensorsToSample() | ||
| } |
72 changes: 72 additions & 0 deletions
72
zoo/src/main/scala/com/intel/analytics/zoo/feature/common/SeqToMultipleTensors.scala
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,72 @@ | ||
| /* | ||
| * Copyright 2018 Analytics Zoo Authors. | ||
| * | ||
| * Licensed under the Apache License, Version 2.0 (the "License"); | ||
| * you may not use this file except in compliance with the License. | ||
| * You may obtain a copy of the License at | ||
| * | ||
| * http://www.apache.org/licenses/LICENSE-2.0 | ||
| * | ||
| * Unless required by applicable law or agreed to in writing, software | ||
| * distributed under the License is distributed on an "AS IS" BASIS, | ||
| * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
| * See the License for the specific language governing permissions and | ||
| * limitations under the License. | ||
| */ | ||
| package com.intel.analytics.zoo.feature.common | ||
|
|
||
| import com.intel.analytics.bigdl.tensor.Tensor | ||
| import com.intel.analytics.bigdl.tensor.TensorNumericMath.TensorNumeric | ||
|
|
||
| import scala.reflect.ClassTag | ||
|
|
||
| /** | ||
| * a Preprocessing that converts Array[Float], Array[Double] or MLlib Vector to multiple tensors | ||
| * for multi-input models | ||
| * @param multiSizes dimensions of target Tensors. | ||
| */ | ||
| class SeqToMultipleTensors[T: ClassTag](multiSizes: Array[Array[Int]]) | ||
| (implicit ev: TensorNumeric[T]) extends Preprocessing[Any, Array[Tensor[T]]] { | ||
|
|
||
| override def apply(prev: Iterator[Any]): Iterator[Array[Tensor[T]]] = { | ||
| prev.map { f => | ||
| val tensors = f match { | ||
| case sd: Seq[Any] => matchSeq(sd) | ||
| case _ => throw new IllegalArgumentException("SeqToTensor only supports Float, Double, " + | ||
| s"Array[Float], Array[Double] or MLlib Vector but got $f") | ||
| } | ||
| tensors | ||
| } | ||
| } | ||
|
|
||
| def matchSeq(list: Seq[Any]): Array[Tensor[T]] = { | ||
| val rawData = list.head match { | ||
| case dd: Double => list.asInstanceOf[Seq[Double]].map(ev.fromType(_)).toArray | ||
| case ff: Float => list.asInstanceOf[Seq[Float]].map(ev.fromType(_)).toArray | ||
| case ii: Int => list.asInstanceOf[Seq[Int]].map(ev.fromType(_)).toArray | ||
| case _ => throw new IllegalArgumentException(s"SeqToTensor only supports Array[Int], " + | ||
| s"Array[Float] and Array[Double] for ArrayType, but got $list") | ||
| } | ||
|
|
||
| require(multiSizes.map(s => s.product).sum == rawData.length, s"feature columns length " + | ||
| s"${rawData.length} does not match with the sum of tensors" + | ||
| s" ${multiSizes.map(a => a.mkString(",")).mkString("\n")}") | ||
|
|
||
| var cur = 0 | ||
| val tensors = multiSizes.map { size => | ||
| val rawLength = size.product | ||
| val t = Tensor(rawData.slice(cur, cur + rawLength), size) | ||
| cur += rawLength | ||
| t | ||
| } | ||
| tensors | ||
| } | ||
| } | ||
|
|
||
|
|
||
| object SeqToMultipleTensors { | ||
| def apply[T: ClassTag]( | ||
| multiSizes: Array[Array[Int]] | ||
| )(implicit ev: TensorNumeric[T]): SeqToMultipleTensors[T] = | ||
| new SeqToMultipleTensors[T](multiSizes) | ||
| } |
Oops, something went wrong.