是否可以在没有CUDA的情况下在H2O中建立深水/ TensorFlow模型

Question

我的目标是在机器上将H2O与没有CUDA的TensorFlow集成。

由于TensorFlow支持CPU和GPU执行，我希望在没有CUDA的情况下实现H2O / TensorFlow集成。但是我在system specifications of Deep Water中提到CUDA软件时感到非常困惑。

我曾试图在H2O Flow中建造Deep Water / TensorFlow模型，但失败了。我执行的步骤：

1. 已下载H2O standalone JAR;
2. 像往常一样在H2O Flow中创建数据框;
3. 试图建立一个分别选择Deep Water和tensorflow作为算法和后端的模型;
4. 得到以下例外：

java.lang.RuntimeException: Unable to initialize the native Deep Learning backend: No backend found. Cannot build a Deep Water model.
    at hex.deepwater.DeepWaterModelInfo.setupNativeBackend(DeepWaterModelInfo.java:246)
    at hex.deepwater.DeepWaterModelInfo.(DeepWaterModelInfo.java:193)
    at hex.deepwater.DeepWaterModel.(DeepWaterModel.java:225)
    at hex.deepwater.DeepWater$DeepWaterDriver.buildModel(DeepWater.java:127)
    at hex.deepwater.DeepWater$DeepWaterDriver.computeImpl(DeepWater.java:114)
    at hex.ModelBuilder$Driver.compute2(ModelBuilder.java:169)
    at hex.deepwater.DeepWater$DeepWaterDriver.compute2(DeepWater.java:107)
    at water.H2O$H2OCountedCompleter.compute(H2O.java:1220)
    at jsr166y.CountedCompleter.exec(CountedCompleter.java:468)
    at jsr166y.ForkJoinTask.doExec(ForkJoinTask.java:263)
    at jsr166y.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:974)
    at jsr166y.ForkJoinPool.runWorker(ForkJoinPool.java:1477)
    at jsr166y.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:104)

所以我的问题是：

1. 是否有可能在没有CUDA的H2O中建立深水/ TensorFlow模型？
2. 如果是，我该怎么办才能让它发挥作用？如果不是，还有其他选项可以在没有CUDA的情况下集成H2O和TensorFlow吗？

更新1 ：

我已将gpu参数设置为false，并尝试使用所有可能的后端再次构建模型。 caffe和tensorflow都产生如上所示的相同堆栈跟踪。 mxnet也失败了，但有两个不同的堆栈跟踪。

mxnet（首次尝试构建模型）：

java.lang.RuntimeException: Unable to initialize the native Deep Learning backend: null
    at hex.deepwater.DeepWaterModelInfo.setupNativeBackend(DeepWaterModelInfo.java:246)
    at hex.deepwater.DeepWaterModelInfo.(DeepWaterModelInfo.java:193)
    at hex.deepwater.DeepWaterModel.(DeepWaterModel.java:225)
    at hex.deepwater.DeepWater$DeepWaterDriver.buildModel(DeepWater.java:127)
    at hex.deepwater.DeepWater$DeepWaterDriver.computeImpl(DeepWater.java:114)
    at hex.ModelBuilder$Driver.compute2(ModelBuilder.java:169)
    at hex.deepwater.DeepWater$DeepWaterDriver.compute2(DeepWater.java:107)
    at water.H2O$H2OCountedCompleter.compute(H2O.java:1220)
    at jsr166y.CountedCompleter.exec(CountedCompleter.java:468)
    at jsr166y.ForkJoinTask.doExec(ForkJoinTask.java:263)
    at jsr166y.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:974)
    at jsr166y.ForkJoinPool.runWorker(ForkJoinPool.java:1477)
    at jsr166y.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:104)

mxnet（后续尝试）：

java.lang.RuntimeException: Unable to initialize the native Deep Learning backend: Could not initialize class deepwater.backends.mxnet.MXNetBackend$MXNetLoader
    at hex.deepwater.DeepWaterModelInfo.setupNativeBackend(DeepWaterModelInfo.java:246)
    at hex.deepwater.DeepWaterModelInfo.(DeepWaterModelInfo.java:193)
    at hex.deepwater.DeepWaterModel.(DeepWaterModel.java:225)
    at hex.deepwater.DeepWater$DeepWaterDriver.buildModel(DeepWater.java:127)
    at hex.deepwater.DeepWater$DeepWaterDriver.computeImpl(DeepWater.java:114)
    at hex.ModelBuilder$Driver.compute2(ModelBuilder.java:169)
    at hex.deepwater.DeepWater$DeepWaterDriver.compute2(DeepWater.java:107)
    at water.H2O$H2OCountedCompleter.compute(H2O.java:1220)
    at jsr166y.CountedCompleter.exec(CountedCompleter.java:468)
    at jsr166y.ForkJoinTask.doExec(ForkJoinTask.java:263)
    at jsr166y.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:974)
    at jsr166y.ForkJoinPool.runWorker(ForkJoinPool.java:1477)
    at jsr166y.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:104)

更新2

环境：

• SW：CentOS Linux版本7.3.1611（核心版），Java HotSpot 64位服务器VM（版本25.121-b13，混合模式）;
• HW：在Xeon CPU E5-2620 v4上运行的虚拟机，具有4个内核和8 GB RAM可用。没有可用的物理GPU，lspci -vnn | grep VGA会返回00:0f.0 VGA compatible controller [0300]: VMware SVGA II Adapter [15ad:0405] (prog-if 00 [VGA controller])

我已清除/ tmp目录并再次尝试使用mxnet。在第一次尝试时，我遇到了新的例外：

java.lang.RuntimeException: Unable to initialize the native Deep Learning backend: /tmp/libmxnet.so: libcudart.so.8.0: cannot open shared object file: No such file or directory
    at hex.deepwater.DeepWaterModelInfo.setupNativeBackend(DeepWaterModelInfo.java:246)
    at hex.deepwater.DeepWaterModelInfo.(DeepWaterModelInfo.java:193)
    at hex.deepwater.DeepWaterModel.(DeepWaterModel.java:225)
    at hex.deepwater.DeepWater$DeepWaterDriver.buildModel(DeepWater.java:127)
    at hex.deepwater.DeepWater$DeepWaterDriver.computeImpl(DeepWater.java:114)
    at hex.ModelBuilder$Driver.compute2(ModelBuilder.java:169)
    at hex.deepwater.DeepWater$DeepWaterDriver.compute2(DeepWater.java:107)
    at water.H2O$H2OCountedCompleter.compute(H2O.java:1220)
    at jsr166y.CountedCompleter.exec(CountedCompleter.java:468)
    at jsr166y.ForkJoinTask.doExec(ForkJoinTask.java:263)
    at jsr166y.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:974)
    at jsr166y.ForkJoinPool.runWorker(ForkJoinPool.java:1477)
    at jsr166y.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:104)

文件/tmp/libmxnet.so已存在，其权限为-rw-rw-r--。

Answer 1

第一个问题的答案如下：

你确定可以在没有GPU的情况下运行深水，它会非常慢。当您使用FLOW时，可以禁用gpu设置，如下所示（默认情况下为TRUE）

您也可以在FLOW单元格中将gpu设置为false，如下所示：

＆＃34; gpu＆＃34;：false

然而，您的主要问题是没有后端（mxnet，tensorflow，caffe）可用于运行您的代码。我们确实用mxnet测试了gpu标志设置。请尝试详细了解上述错误。