metadata

license: bsd
language:
  - tr
metrics:
  - accuracy
pipeline_tag: text-classification

Model Card for Model ID

Turkish Toxic Comment Classifier trained on Turkish comments filtered from Jigsaw Multilingual Toxic Comment Classification. The model weights are initialized with dbmdz/convbert-base-turkish-mc4-uncased

Quick Usage

from transformers import pipeline 
pipe = pipeline("text-classification", model="gokceuludogan/convbert-base-turkish-mc4-toxicity-uncased")
pipe("Ulan sen bizim köyü bizden iyi mi tanıyorsun ki bizim yazdıklarımızı siliyorsun?")

LABEL_1 denotes Toxic while LABEL_0 refers to Neutral.