org.nd4j.linalg.factory.Nd4j.sizeOfDataType()方法的使用及代码示例

x33g5p2x  于2022-01-24 转载在 其他  
字(6.8k)|赞(0)|评价(0)|浏览(120)

本文整理了Java中org.nd4j.linalg.factory.Nd4j.sizeOfDataType()方法的一些代码示例，展示了Nd4j.sizeOfDataType()的具体用法。这些代码示例主要来源于Github/Stackoverflow/Maven等平台，是从一些精选项目中提取出来的代码，具有较强的参考意义，能在一定程度帮忙到你。Nd4j.sizeOfDataType()方法的具体详情如下：
包路径：org.nd4j.linalg.factory.Nd4j
类名称：Nd4j
方法名：sizeOfDataType

Nd4j.sizeOfDataType介绍

[英]This method returns sizeOf(currentDataType), in bytes
[中]此方法返回sizeOf（currentDataType），以字节为单位

代码示例

代码示例来源：origin: deeplearning4j/nd4j

  1. /**
  2.  * This method returns amount of shared memory required for this specific Aggregate.
  3.  * PLEASE NOTE: this method is especially important for CUDA backend. On CPU backend it might be ignored, depending on Aggregate.
  4.  *
  5.  * @return
  6.  */
  7. @Override
  8. public int getSharedMemorySize() {
  9.   return (vectorLength * Nd4j.sizeOfDataType()) + 512;
  10. }

代码示例来源：origin: deeplearning4j/nd4j

  1. @Override
  2. public int getSharedMemorySize() {
  3.   return (vectorLength * Nd4j.sizeOfDataType() * 2) + 512;
  4. }

代码示例来源：origin: deeplearning4j/nd4j

  1. /**
  2.  * This method returns amount of shared memory required for this specific Aggregate.
  3.  * PLEASE NOTE: this method is especially important for CUDA backend. On CPU backend it might be ignored, depending on Aggregate.
  4.  *
  5.  * @return
  6.  */
  7. @Override
  8. public int getSharedMemorySize() {
  9.   return (getThreadsPerInstance() * Nd4j.sizeOfDataType()) + 512;
  10. }

代码示例来源：origin: deeplearning4j/nd4j

  1. /**
  2.  * This method returns amount of shared memory required for this specific Aggregate.
  3.  * PLEASE NOTE: this method is especially important for CUDA backend. On
  4.  * CPU backend it might be ignored, depending on Aggregate.
  5.  *
  6.  * @return
  7.  */
  8. @Override
  9. public int getSharedMemorySize() {
  10.   return (getThreadsPerInstance() * Nd4j.sizeOfDataType()) + 256;
  11. }

代码示例来源：origin: deeplearning4j/nd4j

  1. /**
  2.  * This method returns amount of shared memory required for this specific Aggregate.
  3.  * PLEASE NOTE: this method is especially important for CUDA backend. On CPU backend it might be ignored, depending on Aggregate.
  4.  *
  5.  * @return
  6.  */
  7. @Override
  8. public int getSharedMemorySize() {
  9.   return (getThreadsPerInstance() * Nd4j.sizeOfDataType()) + 512;
  10. }

代码示例来源：origin: deeplearning4j/nd4j

  1. /**
  2.  * This method returns sizeOf(currentDataType), in bytes
  3.  *
  4.  * @return number of bytes per element
  5.  */
  6. public static int sizeOfDataType() {
  7.   return sizeOfDataType(Nd4j.dataType());
  8. }

代码示例来源：origin: deeplearning4j/nd4j

  1. /**
  2.  * This method returns memory used by this DataSet
  3.  *
  4.  * @return
  5.  */
  6. @Override
  7. public long getMemoryFootprint() {
  8.   long reqMem = features.lengthLong() * Nd4j.sizeOfDataType();
  9.   reqMem += labels == null ? 0 : labels.lengthLong() * Nd4j.sizeOfDataType();
  10.   reqMem += featuresMask == null ? 0 : featuresMask.lengthLong() * Nd4j.sizeOfDataType();
  11.   reqMem += labelsMask == null ? 0 : labelsMask.lengthLong() * Nd4j.sizeOfDataType();
  12.   return reqMem;
  13. }

代码示例来源：origin: deeplearning4j/nd4j

  1. /**
  2.  * This method returns memory used by this DataSet
  3.  *
  4.  * @return
  5.  */
  6. @Override
  7. public long getMemoryFootprint() {
  8.   long reqMem = 0;
  9.   for (INDArray f : features)
  10.     reqMem += f == null ? 0 : f.lengthLong() * Nd4j.sizeOfDataType();
  11.   if (featuresMaskArrays != null)
  12.     for (INDArray f : featuresMaskArrays)
  13.       reqMem += f == null ? 0 : f.lengthLong() * Nd4j.sizeOfDataType();
  14.   if (labelsMaskArrays != null)
  15.     for (INDArray f : labelsMaskArrays)
  16.       reqMem += f == null ? 0 : f.lengthLong() * Nd4j.sizeOfDataType();
  17.   if (labels != null)
  18.     for (INDArray f : labels)
  19.       reqMem += f == null ? 0 : f.lengthLong() * Nd4j.sizeOfDataType();
  20.   return reqMem;
  21. }

代码示例来源：origin: deeplearning4j/nd4j

  1. requiredMemory += div;
  2. long numElements = requiredMemory / Nd4j.sizeOfDataType(type);

代码示例来源：origin: deeplearning4j/nd4j

  1. protected static DataBuffer internalCreateBuffer(double[] data) {
  2.   val perfX = PerformanceTracker.getInstance().helperStartTransaction();
  3.   val buffer = Nd4j.createBuffer(data);
  4.   PerformanceTracker.getInstance().helperRegisterTransaction(0, perfX, data.length * Nd4j.sizeOfDataType(), MemcpyDirection.HOST_TO_HOST);
  5.   return buffer;
  6. }

代码示例来源：origin: deeplearning4j/nd4j

  1. protected static DataBuffer internalCreateBuffer(float[] data, long offset) {
  2.   val perfX = PerformanceTracker.getInstance().helperStartTransaction();
  3.   val buffer = Nd4j.createBuffer(data, offset);
  4.   PerformanceTracker.getInstance().helperRegisterTransaction(0, perfX, data.length * Nd4j.sizeOfDataType(), MemcpyDirection.HOST_TO_HOST);
  5.   return buffer;
  6. }

代码示例来源：origin: deeplearning4j/nd4j

  1. protected static DataBuffer internalCreateBuffer(double[] data, long offset) {
  2.   val perfX = PerformanceTracker.getInstance().helperStartTransaction();
  3.   val buffer = Nd4j.createBuffer(data, offset);
  4.   PerformanceTracker.getInstance().helperRegisterTransaction(0, perfX, data.length * Nd4j.sizeOfDataType(), MemcpyDirection.HOST_TO_HOST);
  5.   return buffer;
  6. }

代码示例来源：origin: deeplearning4j/nd4j

  1. protected static DataBuffer internalCreateBuffer(int[] data, long offset) {
  2.   val perfX = PerformanceTracker.getInstance().helperStartTransaction();
  3.   val buffer = Nd4j.createBuffer(data, offset);
  4.   PerformanceTracker.getInstance().helperRegisterTransaction(0, perfX, data.length * Nd4j.sizeOfDataType(), MemcpyDirection.HOST_TO_HOST);
  5.   return buffer;
  6. }

代码示例来源：origin: deeplearning4j/nd4j

  1. protected static DataBuffer internalCreateBuffer(float[] data) {
  2.   val perfX = PerformanceTracker.getInstance().helperStartTransaction();
  3.   val buffer = Nd4j.createBuffer(data);
  4.   PerformanceTracker.getInstance().helperRegisterTransaction(0, perfX, data.length * Nd4j.sizeOfDataType(), MemcpyDirection.HOST_TO_HOST);
  5.   return buffer;
  6. }

代码示例来源：origin: deeplearning4j/nd4j

  1. protected static DataBuffer internalCreateBuffer(int[] data) {
  2.   val perfX = PerformanceTracker.getInstance().helperStartTransaction();
  3.   val buffer = Nd4j.createBuffer(data);
  4.   PerformanceTracker.getInstance().helperRegisterTransaction(0, perfX, data.length * Nd4j.sizeOfDataType(), MemcpyDirection.HOST_TO_HOST);
  5.   return buffer;
  6. }

代码示例来源：origin: deeplearning4j/nd4j

  1. /**
  2.  *
  3.  * @param data
  4.  * @param shape
  5.  * @param stride
  6.  * @param offset
  7.  * @param ordering
  8.  */
  9. public BaseNDArray(float[] data, int[] shape, int[] stride, long offset, char ordering) {
  10.   setShapeInformation(Nd4j.getShapeInfoProvider().createShapeInformation(shape, stride, offset,
  11.       Shape.elementWiseStride(shape, stride, ordering == 'f'), ordering));
  12.   if (data != null && data.length > 0) {
  13.     val perfD = PerformanceTracker.getInstance().helperStartTransaction();
  14.     this.data = internalCreateBuffer(data, offset);
  15.     PerformanceTracker.getInstance().helperRegisterTransaction(0, perfD, data.length * Nd4j.sizeOfDataType(), MemcpyDirection.HOST_TO_HOST);
  16.     if (offset >= data.length)
  17.       throw new IllegalArgumentException("invalid offset: must be < data.length");
  18.   }
  19.   init(shape, stride);
  20. }

代码示例来源：origin: org.nd4j/nd4j-api

  1. @Override
  2. public int getSharedMemorySize() {
  3.   return (vectorLength * Nd4j.sizeOfDataType() * 2) + 512;
  4. }

代码示例来源：origin: org.nd4j/nd4j-api

  1. /**
  2.  * This method returns amount of shared memory required for this specific Aggregate.
  3.  * PLEASE NOTE: this method is especially important for CUDA backend. On CPU backend it might be ignored, depending on Aggregate.
  4.  *
  5.  * @return
  6.  */
  7. @Override
  8. public int getSharedMemorySize() {
  9.   return (getThreadsPerInstance() * Nd4j.sizeOfDataType()) + 512;
  10. }

代码示例来源：origin: org.nd4j/nd4j-api

  1. /**
  2.  * This method returns sizeOf(currentDataType), in bytes
  3.  *
  4.  * @return number of bytes per element
  5.  */
  6. public static int sizeOfDataType() {
  7.   return sizeOfDataType(Nd4j.dataType());
  8. }

代码示例来源：origin: org.nd4j/nd4j-api

  1. /**
  2.  * This method returns amount of shared memory required for this specific Aggregate.
  3.  * PLEASE NOTE: this method is especially important for CUDA backend. On
  4.  * CPU backend it might be ignored, depending on Aggregate.
  5.  *
  6.  * @return
  7.  */
  8. @Override
  9. public int getSharedMemorySize() {
  10.   return (getThreadsPerInstance() * Nd4j.sizeOfDataType()) + 256;
  11. }

内容来源于网络，如有侵权，请联系作者删除！

Nd4j

相关文章

    查看更多

    热门标签

    更多
    JavaquerypythonNode开发语言requestUtil数据库Table后端算法LoggerMessageElementParser

    最新文章

    更多

    Nd4j类方法

    • 蜀ICP备13028337号-1 大数据知识库 https://www.saoniuhuo.com © All rights reserved
    • 本站内容来源互联网,如果侵犯您的权益请联系我们删除, 联系方式：448109455@qq.com