h2o | 易学教程

线程协作工具之Semaphore、CountDownLatch、CyclicBarrier

阅读更多关于线程协作工具之Semaphore、CountDownLatch、CyclicBarrier

【推荐】2019 Java 开发者跳槽指南.pdf(吐血整理) >>> 各种线程协作工具常见线程协作工具：读写锁ReadWriteLock 信号量Semaphore 倒计时门栓CountDownLatch 循环栅栏CyclicBarrier 线程本地变量ThreadLocal 主要讲信号量Semaphore，倒计时门栓CountDownLatch，循环栅栏CyclicBarrier 然后根据两个LeetCode题目来应用一下。第一题：三个不同的线程将会共用一个 Foo 实例。线程 A 将会调用 one() 方法线程 B 将会调用 two() 方法线程 C 将会调用 three() 方法请设计修改程序，以确保 two() 方法在 one() 方法之后被执行，three() 方法在 two() 方法之后被执行。原生解法应用场景及用法基于notify/wait，所有的线程间通信类似通知的机制本质上都是notifyAll，多能用这个来实现。 class Foo { private static AtomicInteger flag = new AtomicInteger(0); private static Object Lock1 = new Object(); public Foo() { } public void first(Runnable

Using MOJOS in H2O Steam Prediction Service Builder

阅读更多关于 Using MOJOS in H2O Steam Prediction Service Builder

问题 In h2o 's documentation for the Steam Prediction Service Builder, here, it says that the service builder can compile both h2o pojo s (.java files) and mojo s (downloaded from h2o flow in my case as a .zip (version 3.10.5.2), which I have been using in the manner shown here). However, doing something like this: gives this error: Problem accessing /makewar. Reason: Compilation of pojo failed exit value 1 warning: [options] bootstrap class path not set in conjunction with -source 1.6 error:

How to directly plot ROC of h2o model object in R

阅读更多关于 How to directly plot ROC of h2o model object in R

问题 My apologies if I'm missing something obvious. I've been thoroughly enjoying working with h2o in the last few days using R interface. I would like to evaluate my model, say a random forest, by plotting an ROC. The documentation seems to suggest that there is a straightforward way to do that: Interpreting a DRF Model By default, the following output displays: Model parameters (hidden) A graph of the scoring history (number of trees vs. training MSE) A graph of the ROC curve (TPR vs. FPR) A

Improve h2o DRF runtime on a multi-node cluster

阅读更多关于 Improve h2o DRF runtime on a multi-node cluster

问题 I am currently running h2o 's DRF algorithm an a 3-node EC2 cluster (the h2o server spans across all 3 nodes). My data set has 1m rows and 41 columns (40 predictors and 1 response). I use the R bindings to control the cluster and the RF call is as follows model=h2o.randomForest(x=x, y=y, ignore_const_cols=TRUE, training_frame=train_data, seed=1234, mtries=7, ntrees=2000, max_depth=15, min_rows=50, stopping_rounds=3, stopping_metric="MSE", stopping_tolerance=2e-5) For the 3-node cluster (c4

Implementing custom stopping metrics to optimize during training in H2O model directly from R

阅读更多关于 Implementing custom stopping metrics to optimize during training in H2O model directly from R

问题 I'm trying to implement the FBeta_Score() of the MLmetrics R package: FBeta_Score <- function(y_true, y_pred, positive = NULL, beta = 1) { Confusion_DF <- ConfusionDF(y_pred, y_true) if (is.null(positive) == TRUE) positive <- as.character(Confusion_DF[1,1]) Precision <- Precision(y_true, y_pred, positive) Recall <- Recall(y_true, y_pred, positive) Fbeta_Score <- (1 + beta^2) * (Precision * Recall) / (beta^2 * Precision + Recall) return(Fbeta_Score) } in the H2O distributed random forest model

Error when starting h2o.init() in R

阅读更多关于 Error when starting h2o.init() in R

问题 I am getting an error when trying to start h2o. Running the following start-up commands: library(h2o) localH2O = h2o.init(ip = "localhost", port = 54321, startH2O = TRUE) Gives the following error: Error in system2(command, "-version", stdout = TRUE, stderr = TRUE) : '""' not found It's not particularly enlightning. Does anyone have a hint as to what is the problem? Java is installed on the system, but unlike the question here I do not get any Java warnings. Running system("java -version")

h2o Cluster total memory unexpectedly low

阅读更多关于 h2o Cluster total memory unexpectedly low

问题 I'm working on a remote server with 196 GB of ram but when I use h2o.init(nthreads = -1) it says that the total cluster memory is .96 GB. Furthermore, when I try to use max_mem_size = "2g" I get an error saying that the JVM could not be created. Now, the server is using a 32 bit Java version and I do not have admin access. What do I need to do to get more cluster memory? Thanks 回答1: Now, the server is using a 32 bit Java version and I do not have admin access. What do I need to do to get more

h2o deeplearning error when specifying nfolds for cross validation

阅读更多关于 h2o deeplearning error when specifying nfolds for cross validation

问题 has this issue been resolved by now? I encounter the same error message. Usecase: I am doing binary classification using h2o's deeplearning() function. Below, I provide randomly generated data the same size as my actual usecase. System specs: # R version 3.3.2 (2016-10-31) # Platform: x86_64-w64-mingw32/x64 (64-bit) # Running under: Windows >= 8 x64 (build 9200) # h2o version h2o_3.20.0.2 I am currently learning how to use h2o, so I have played with that function quite a bit. Everything runs

How to visualize H2O Tree?

阅读更多关于 How to visualize H2O Tree?

问题 I have df data_categorical and a model model . I converted my df to h2o frame with data = h2o.H2OFrame(data_categorical) and trained my model with model = H2ORandomForestEstimator(ntrees=1, max_depth=20, nfolds=10) # Train model model.train(x=training_columns, y=response_column, training_frame=train) I'm trying to visualize the tree that is created (note that I only need one tree) but I can't seem to do that. I downloaded the mojo file with model.download_mojo(path,get_genmodel_jar=True) But

h2o model list dtypes for each feature

阅读更多关于 h2o model list dtypes for each feature

问题 Is there a way to list the expected data type for each feature for an existing h2o model, without looking at the training data? Similar to the output of h2o_frame.types, but without looking at the training data just looking at the model object. In this case I have a saved model file (the generic type that can be imported back into h2o, not mojo/pojo). The algorithm is gbm. 回答1: This functionality is not currently available in H2O, but I've created a jira ticket to request it as a new feature,