Create a Task model to aggregate sets of data used for training and testing models about mlplayground

[Abhit03]

Regarding the ManyToMany relationship. It makes sense to me.
Since as you mentioned - We will have a table which will point to the foreign keys of both Task and Histogram. And the Task fields will not have to appear in the Histogram models.

[nothingface0]

From @XavierAtCERN:

what would be good for the front end would be to have a /create view with the three options I mentioned:

• train on first lumisections of a run, test on the following
  • Form with run number and start / stop lumisection for train + start / stop lumisection for test
• train on a few runs, test on the following
  • Form with run numbers to train on and run number(s) to test
• user defined train/test split
  • Drag and drop area for a two column file (runs+lumisections) for train and one for test