atasoglu/turkish-tiny-bert-uncased-offenseval2020_tr

This is a lightweight offensive language detection model fine-tuned with coltekin/offenseval2020_tr dataset on ytu-ce-cosmos/turkish-tiny-bert-uncased.

Usage

from transformers import pipeline
pipe = pipeline("text-classification", "atasoglu/turkish-tiny-bert-uncased-offenseval2020_tr")
print(pipe("bu bir test metnidir.", top_k=None))
# [{'label': 'NOT', 'score': 0.9964759945869446}, {'label': 'OFF', 'score': 0.0035239490680396557}]

Test Results

Test results examined on the test split of fine-tuning dataset.

precision recall f1-score support
NOT 0.8966 0.9317 0.9138 2812
OFF 0.6832 0.5782 0.6263 716
accuracy 0.8600 3528
macro avg 0.7899 0.7550 0.7701 3528
weighted avg 0.8533 0.8600 0.8555 3528
Downloads last month
13
Safetensors
Model size
4.58M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Dataset used to train atasoglu/turkish-tiny-bert-uncased-offenseval2020_tr