Modelpublic

DeadBeast/mbert-base-cased-finetuned-bengali-fakenews

sourceHugging Faceapache-2.0updated 5y agoView on Hugging Face

3likes31downloads

Model Card

mBERT-base-cased-finetuned-bengali-fakenews

This model is a fine-tune checkpoint of mBERT-base-cased over [Bengali-fake-news Dataset](https://www.kaggle.com/cryptexcode/banfakenews) for Text classification. This model reaches an accuracy of 96.3 with an f1-score of 79.1 on the dev set.

How to use?

Task: binary-classification

—LABEL_1: Authentic (Authentic means news is authentic)
—LABEL_0: Fake (Fake means news is fake)

from transformers import pipeline
print(pipeline("sentiment-analysis",model="DeadBeast/mbert-base-cased-finetuned-bengali-fakenews",tokenizer="DeadBeast/mbert-base-cased-finetuned-bengali-fakenews")("অভিনেতা আফজাল শরীফকে ২০ লাখ টাকার অনুদান অসুস্থ অভিনেতা আফজাল শরীফকে চিকিৎসার জন্য ২০ লাখ টাকা অনুদান দিয়েছেন প্রধানমন্ত্রী শেখ হাসিনা।"))