Larger Input Data Set

[asim215]

Hello. I curious to ask following.
If we put much larger data sets as input to train model, is it possible to get better final result?
For example, more code (github), books (gutenberg).

[binarybottle]

Please see #86 (comment) for an announcement of the new Engram approach/layouts/repo! I think you will appreciate the data-driven approach.