Skip to content

Latest commit

 

History

History
21 lines (15 loc) · 759 Bytes

README.md

File metadata and controls

21 lines (15 loc) · 759 Bytes

NepaliSentiment

Nepali Sentiment Classification.

BAD INPUT BAD OUTPUT .(As data is collected from scraping web pages and not been preprocessed, so kindly clean your self) Model

Score

This score are generated without cleaning Data.(On RAW DATA).

Model Score
TF-IDF + SVC 0.55
TF-IDF + Emoji Decoded 0.59
DeBerta large Mnli (Feature Extraction) 0.45
DeBerta Base Nepali (Feature Extraction) 0.63
NepaliBERT (fine tunning) 0.746(F1:- 0.759, 0.840, 0.612)
NepBERTa (fine tunning) 0.741(F1:-0.749, 0.840, 0.619

Some of the code are on private repo.(If you are interested then you can ask me for code).(Need To Be Done Some Cleaning)