kuhy / glade

GLADE is a tool for automatically synthesizing program input grammars
Apache License 2.0
6 stars 1 forks source link

[Question] Runtime estimation or cutoff #1

Open stheid opened 3 years ago

stheid commented 3 years ago

Hi,

I hope its ok if i ask a technical question here:

In our project we are trying to use the learned grammar, hence we have not yet worked a lot with the theory on grammar learning, but we where wondering what influences the runtime of the grammar learning. As an experiment i started with learning the grammar of png files. I assume thats a little more challenging than xml but i wonder it there are ways to cutoff the grammar inference and have a look at the intermediate model. Also i am wondering if we are talking about trainingtime of weeks or hours. in this case.

Currently i have let it train for 16h and it still says "learning grammar"

Best, Stefan

stheid commented 3 years ago

To be a little more precise: As we are also working in a fuzzing scenario we are not so much interested in a very precise model, however we need to train models for various domains.

McSinyx commented 2 years ago

trainingtime of weeks or hours

Did it finish? If so, how long did it take?

stheid commented 2 years ago

No, we killed it and never really achieved to train it. i think the tool was not suitable for our case