Hidden

NLP course. Text classification: toxic comments

Your task is to classify toxic comments into 6 classes (from 0 to 5). The metric is accuracy, the baselines quality is 0.60 and 0.7. If you beat the first baseline you will get 12 points, if you bet the second baseline you will get 25 points. If you will be the first in your group, you'll get 3 bonus points.
The test data is a set of strings, one comment per line. You need to return a label for each comment, also one per line.

The baseline solution is here.