Toxic comments are a big problem for society these days, some people are very vulnerable and, after reading negative reviews, they can become depressed and stop visiting the product in which they were insulted. Therefore, the organizers of Tinkoff decided to offer this problem at the hackathon at MIPT on April 14, 2021. This implementation of detecting toxic comments is written in basic ruBert, trained on Russian Wikipedia, on top of which there are 2 fully connected layers (fine tuning applied). On the test sample, the algorithm worked quite well, given the complexity of the toxic vocabulary. Macro Average score f1 showed 0.76.
-
Notifications
You must be signed in to change notification settings - Fork 0
Varfalamei/Toxic_detection_bert
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
No description, website, or topics provided.
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published