"Error executing Deep Learning"

naveen_bharadwanaveen_bharadwa Member Posts: 9 Contributor I
edited June 2019 in Help

Hi,

 

I've been trying to build a model using the deep learning operator. My dataset has two columns both are text in type and one of them is the attribute and the other one is the label. When I feed this dataset into RapidMiner, it fails at the Deep Learning operator saying "Error while executing the H2O model: {0}"

 

Attached is the error message and the same thing is seen in the logs. No additional help there.

 

 

h2oModel.png

 

Any help would be appreciated. Is there a problem with the way the data is being fed and is there a format other than what I am doing here?

 

Regards,

Naveen

 

 

 

Answers

  • Thomas_OttThomas_Ott RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,761 Unicorn

    Without looking at your RMP sans data, my first question is did you tokenize your text data? I'm assuming you want to do some sort of sentiment analysis. This means you will need to use the Text Processing extension. The Deep Learning learner can't just read text data without some sort of transformation.

     

    Also, I would not just train the deep learning model without knowing what the performance will be, so I would also use a Cross Validation. 

  • naveen_bharadwanaveen_bharadwa Member Posts: 9 Contributor I

    Hi,

     

    The data transformation part is not in the RM process. I did transform the data and while reading the CSV, I made sure that I'm only reading the last two columns as seen in the image attached below. This is basically just the text and its label.

    Screenshot from 2018-02-19 18-03-46.png

     

    Regards,

    Naveen

     

  • lionelderkrikorlionelderkrikor Moderator, RapidMiner Certified Analyst, Member Posts: 1,195 Unicorn

    Hi @naveen_bharadwa,

     

    I created an excel file with the last two columns as seen in the image you shown and it works with the Deep learning model for me.

    However 2 elements : 

     - I don't see the Set Role operator in your process. How did you define your label ? set in the Read CSV operator ?

     - I don't see the Select Attributes operator in your process. How are you sure that only the 2 last columns are selected ? set in the Read CSV operator ?

    Maybe can you share your csv file so that we can reproduce your error ?

     

    Regards, 

     

    Lionel

     

  • naveen_bharadwanaveen_bharadwa Member Posts: 9 Contributor I

    Hey,

     

    You can use the import wizard to set the role of the column and also choose which rows you would want to take forward. It's a very handy function RapidMiner has. 

     

    I really appreciate the interest there, but I won't be able to forward the dataset for two main reasons.

    1. The dataset is 3GB in size

    2. I am working on confidential data that shouldn't be shared. 

     

    I can assure you that the description of the data that I've given is apt. Any suggestion?

     

    Regards,

    Naveen

  • lionelderkrikorlionelderkrikor Moderator, RapidMiner Certified Analyst, Member Posts: 1,195 Unicorn

    Hi @naveen_bharadwa,

     

    Since the process works for me, with an extract of your dataset, I would try to update RapidMiner 

    and check the compatibility of your operators.

     

    Best regards, 

     

     

    Lionel

     

     

  • naveen_bharadwanaveen_bharadwa Member Posts: 9 Contributor I

    Hey @lionelderkrikor

     

    Do you think you can share your dataset? That would be a good help. I can reverse engineer my dataset.

     

    Regards,

    Naveen

  • lionelderkrikorlionelderkrikor Moderator, RapidMiner Certified Analyst, Member Posts: 1,195 Unicorn

    Hi @naveen_bharadwa,

     

    You can find my dataset following this link : 

    https://drive.google.com/open?id=1xGIi4I5cIaH7Op4v1i8O0n0Ac_uH1-Bi

     

    It's an excel file, so you have to put an Read Excel operator in entry.

     

    I hope it helps

     

    Regards

     

    Lionel

Sign In or Register to comment.