search

TromboneDavies / PolarOps

0 stars 0 forks source link

Actually assemble the for-reals training data #75

Open divilian opened 2 years ago

divilian commented 2 years ago
  1. Do something like choose a random X threads from each subreddit (X=500?) and put it all in a single .csv called "consider_for_bootstrapping.csv"
  2. Run bootstrap.py with that .csv file.
  3. This will create hand_tagged_plus.csv, which is the "for-reals training data."
divilian commented 2 years ago

(Obviously, this can't be finalized until #73 and #74 are finished.)