dsisds / jforests

Automatically exported from code.google.com/p/jforests
0 stars 0 forks source link

jforests crashed with sample-training-data #2

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago
What steps will reproduce the problem?
1. convert sample-ranking-data to binary format
2. train with sample-ranking-data using sample-ranking-config.properties

What is the expected output? What do you see instead?
jforests should work with given sample-training-data, but it crashed instead 
(see output below)

What version of the product are you using? On what operating system?
jforests-0.3.jar on centOS6.4

Please provide any additional information below.

[resources]$ java -jar jforests-0.3.jar --cmd=generate-bin --ranking --folder . 
--file train.txt
Generating binary files for ranking data sets...
Processing: ./train.txt
10000
20000
30000
40000
Loading values...  [Done in: 0 seconds.]
Making distributions...  [Done in: 0 seconds.]
Making bins...
Feature: 0, type: SHORT
Feature: 1, type: SHORT
Feature: 2, type: BYTE
Feature: 3, type: BYTE
Feature: 4, type: SHORT
Feature: 5, type: NULL
Feature: 6, type: NULL
Feature: 7, type: NULL
Feature: 8, type: NULL
Feature: 9, type: NULL
Feature: 10, type: SHORT
Feature: 11, type: SHORT
Feature: 12, type: SHORT
Feature: 13, type: SHORT
Feature: 14, type: SHORT
Feature: 15, type: SHORT
Feature: 16, type: SHORT
Feature: 17, type: SHORT
Feature: 18, type: SHORT
Feature: 19, type: SHORT
Feature: 20, type: SHORT
Feature: 21, type: SHORT
Feature: 22, type: SHORT
Feature: 23, type: SHORT
Feature: 24, type: SHORT
Feature: 25, type: SHORT
Feature: 26, type: SHORT
Feature: 27, type: SHORT
Feature: 28, type: SHORT
Feature: 29, type: SHORT
Feature: 30, type: SHORT
Feature: 31, type: SHORT
Feature: 32, type: SHORT
Feature: 33, type: SHORT
Feature: 34, type: SHORT
Feature: 35, type: SHORT
Feature: 36, type: SHORT
Feature: 37, type: SHORT
Feature: 38, type: SHORT
Feature: 39, type: SHORT
Feature: 40, type: SHORT
Feature: 41, type: SHORT
Feature: 42, type: SHORT
Feature: 43, type: BYTE
Feature: 44, type: SHORT
Feature: 45, type: BIT
  [Done in: 0 seconds.]
Making features...  [Done in: 0 seconds.]
Creating bin file...  [Done in: 0 seconds.]
[resources]$ java -jar ~/rank/jforests-0.3.jar --cmd=train --ranking 
--config-file sample-ranking-config.properties --train-file train.bin 
--output-model ensemble.txt
Loading datasets...
Finished loading datasets.
java.lang.NullPointerException
    at edu.uci.jforests.learning.boosting.LambdaMART.preprocess(LambdaMART.java:122)
    at edu.uci.jforests.learning.boosting.GradientBoosting.learn(GradientBoosting.java:97)
    at edu.uci.jforests.applications.ClassificationApp.run(ClassificationApp.java:244)
    at edu.uci.jforests.applications.Runner.train(Runner.java:100)
    at edu.uci.jforests.applications.Runner.main(Runner.java:247)
[resources]$ java -version
java version "1.7.0_25"
Java(TM) SE Runtime Environment (build 1.7.0_25-b15)
Java HotSpot(TM) 64-Bit Server VM (build 23.25-b01, mixed mode)
[resources]$ uname -a
Linux centos 2.6.32-358.23.2.el6.x86_64 #1 SMP Wed Oct 16 18:37:12 UTC 2013 
x86_64 x86_64 x86_64 GNU/Linux

Original issue reported on code.google.com by zha...@gmail.com on 5 Nov 2013 at 2:28

GoogleCodeExporter commented 9 years ago
The problem is the lack of validation data. You need to use validation data for 
LambdaMART. 

Original comment by craig.ma...@gmail.com on 23 Apr 2014 at 8:09