You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@systemml.apache.org by "Niketan Pansare (JIRA)" <ji...@apache.org> on 2016/05/27 20:57:12 UTC

[jira] [Created] (SYSTEMML-723) Discuss and agree on the input data format for Deep Learning

Niketan Pansare created SYSTEMML-723:
----------------------------------------

             Summary: Discuss and agree on the input data format for Deep Learning
                 Key: SYSTEMML-723
                 URL: https://issues.apache.org/jira/browse/SYSTEMML-723
             Project: SystemML
          Issue Type: Task
            Reporter: Niketan Pansare


Examples of issues to be discussed:
1. Whether the labels need to be one-based or zero-based.
2. Should one apply z-scoring etc.
3. Input dataset and input labels needs to be seperated in two files and user provides the path to the directory.

As a part of this task, we can provide simple wrappers that allows the user to either generate or download frequently used datasets, transform it into the format needed by SystemML.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)