Saving h2o model in a Windows machine

282 views
Skip to first unread message

Sahiba Chopra

unread,
Feb 11, 2016, 1:18:55 AM2/11/16
to H2O Open Source Scalable Machine Learning - h2ostream
How can I save an h2o model in my Windows machine? Keep getting an "unused argument" error for the file path.

ccl...@gmail.com

unread,
Feb 12, 2016, 12:22:50 PM2/12/16
to H2O Open Source Scalable Machine Learning - h2ostream
On Wednesday, February 10, 2016 at 10:18:55 PM UTC-8, Sahiba Chopra wrote:
> How can I save an h2o model in my Windows machine? Keep getting an "unused argument" error for the file path.

Can you provide more information?
Especially - the version of H2O, the file path you are using, and the full error message dump.
Thanks
Cliff

Todd Morrison

unread,
Feb 13, 2016, 9:43:59 AM2/13/16
to H2O Open Source Scalable Machine Learning - h2ostream
I save models on Windows Server and Windows 10 all the time. This must have something to do with your specific model/config.


On Wednesday, February 10, 2016 at 10:18:55 PM UTC-8, Sahiba Chopra wrote:

Sahiba Chopra

unread,
Feb 15, 2016, 1:44:02 AM2/15/16
to Todd Morrison, H2O Open Source Scalable Machine Learning - h2ostream
I understand that it works on your system but I can't figure out the problem on mine. I've consulted forums but no luck. 

I'm using H2O v 3.6.0.8 and here are the 2 errors that I'm getting:

# Error1###########################

 

> model_path <- h2o.saveModel(object = prostate.dl,path = "D:/Solar_Forecast/Solar Data/Solar_Data_deep_learning/New_Data/MMF_NN_Model",force = FALSE)

 

ERROR: Unexpected HTTP Status code: 400 Bad Request (url = http://127.0.0.1:54321/99/Models.bin/DeepLearning_model_R_1455517614761_5?dir=D%3A%2FSolar_Forecast%2FSolar%20Data%2FSolar_Data_deep_learning%2FNew_Data%2FMMF_NN_Model%2FDeepLearning_model_R_1455517614761_5&force=FALSE)

 

java.lang.IllegalArgumentException

[1] "java.net.URI.create(URI.java:859)"                                                   

 [2] "water.util.FileUtils.getURI(FileUtils.java:44)"                                      

 [3] "water.api.ModelsHandler.exportModel(ModelsHandler.java:227)"                         

 [4] "sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)"                         

 [5] "sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)"       

 [6] "sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)"

[7] "java.lang.reflect.Method.invoke(Method.java:606)"                                    

 [8] "water.api.Handler.handle(Handler.java:64)"                                            

 [9] "water.api.RequestServer.handle(RequestServer.java:644)"                              

[10] "water.api.RequestServer.serve(RequestServer.java:585)"                               

[11] "water.JettyHTTPD$H2oDefaultServlet.doGeneric(JettyHTTPD.java:617)"                   

[12] "water.JettyHTTPD$H2oDefaultServlet.doGet(JettyHTTPD.java:559)"                       

[13] "javax.servlet.http.HttpServlet.service(HttpServlet.java:735)"                        

[14] "javax.servlet.http.HttpServlet.service(HttpServlet.java:848)"                        

[15] "org.eclipse.jetty.servlet.ServletHolder.handle(ServletHolder.java:684)"              

 

Error in .h2o.doSafeREST(h2oRestApiVersion = h2oRestApiVersion, urlSuffix = page,  :

  Illegal character in path at index 23: D:/Solar_Forecast/Solar Data/Solar_Data_deep_learning/New_Data/MMF_NN_Model/DeepLearning_model_R_1455517614761_5

>

#Error2###############################

 

> model_path <- h2o.saveModel(object = prostate.dl)

 

ERROR: Unexpected HTTP Status code: 500 Server Error (url = http://127.0.0.1:54321/99/Models.bin/DeepLearning_model_R_1455517614761_5?dir=%2FDeepLearning_model_R_1455517614761_5&force=FALSE)

 

water.api.FSIOException

[1] "water.persist.PersistFS.create(PersistFS.java:122)"                                  

 [2] "water.serial.KeyedBinarySerializer.save(KeyedBinarySerializer.java:51)"              

 [3] "water.serial.KeyedBinarySerializer.save(KeyedBinarySerializer.java:25)"              

 [4] "water.serial.ObjectTreeBinarySerializer.save(ObjectTreeBinarySerializer.java:93)"    

 [5] "water.api.ModelsHandler.exportModel(ModelsHandler.java:228)"                         

 [6] "sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)"                          

 [7] "sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)"       

 [8] "sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)"

[9] "java.lang.reflect.Method.invoke(Method.java:606)"                                     

[10] "water.api.Handler.handle(Handler.java:64)"                                           

[11] "water.api.RequestServer.handle(RequestServer.java:644)"                              

[12] "water.api.RequestServer.serve(RequestServer.java:585)"                               

[13] "water.JettyHTTPD$H2oDefaultServlet.doGeneric(JettyHTTPD.java:617)"                   

[14] "water.JettyHTTPD$H2oDefaultServlet.doGet(JettyHTTPD.java:559)"                       

[15] "javax.servlet.http.HttpServlet.service(HttpServlet.java:735)"                        

[16] "javax.servlet.http.HttpServlet.service(HttpServlet.java:848)"                        

[17] "org.eclipse.jetty.servlet.ServletHolder.handle(ServletHolder.java:684)"              

 

Error in .h2o.doSafeREST(h2oRestApiVersion = h2oRestApiVersion, urlSuffix = page,  :

  FS IO Failure:

 accessed path : file:/C:/DeepLearning_model_R_1455517614761_5/modelmetrics_DeepLearning_model_R_1455517614761_5_2297637023164670336_on_Key_Frame__C__Users_23163152_AppData_Local_Temp_RtmpuMBNr6_file1f4c6dd44662_csv_1.hex_2.DeepLearning_model_R_1455517614761_5.temporary.train.chunks16.sample.7.74__-5734083892598788322.bin





--
You received this message because you are subscribed to a topic in the Google Groups "H2O Open Source Scalable Machine Learning - h2ostream" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/h2ostream/Bt-dJXNKjDU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to h2ostream+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



--

Tom Kraljevic

unread,
Feb 15, 2016, 11:14:42 AM2/15/16
to Sahiba Chopra, Todd Morrison, H2O Open Source Scalable Machine Learning - h2ostream

Hi,

A new version of h2o was just released last week (3.8.0.3).
This may have some windows fixes.

Otherwise, I suggest trying as simple a path as possible, with no spaces or drive letters, and see if that helps.

Tom


You received this message because you are subscribed to the Google Groups "H2O Open Source Scalable Machine Learning - h2ostream" group.
To unsubscribe from this group and stop receiving emails from it, send an email to h2ostream+...@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages