constantinelignos / morsel Goto Github PK

Hi Constantine,

Thanks for your work on MORSEL. I am testing it out for a research project at my lab, and I noticed that there seems to be an overflow error in the code.

Here is what I see when I start it up with my lexicon:

Loading words...
2583018 types loaded...
-1910967574 tokens loaded.
Init time: 45.85s

Lexicon stats:
Types: 2583018 Tokens: -1910967574

Handling hyphenation...
Lexicon stats:
Types: 2583018 Tokens: -1910967574
Memory status: 3657MB Used, 23501MB Remaining

Starting learning...
Iteration 1
Lexicon stats:
Types: 2583018 Tokens: -1910967574
Base size: 0
Derived size: 0
Unmodeled size: 2583018
Hypothesizing and scoring transforms...
Selecting a transform...

I am not sure if this overflow is going to mess up anything else, but I thought you might want to know. I think if you used a 64 bit integer instead of a 32 bit integer, it might avoid this problem.

Thanks,

Cyrus Shaoul