FiinGroup/phobert-finetuned

FiinGroup-phobert-based for financial news sentiment analysis

This is a PhoBert-base model fine-tuned on ~ 15,000 Vietnamese financial news from Jan 2020 to Dec 2024.

Labels: 0 -> Negative, 1 -> Neutral, 2 -> Positive

Accuracy: 0.903

Example Pipeline

Please don't forget to install transformers library if you have not done it before

!pip install transformers

from transformers import pipeline

model_path = "FiinGroup/phobert-finetuned"
sentiment_task = pipeline("sentiment-analysis", model=model_path, tokenizer=model_path)
sentence = "Covid cases are increasing fast!"
sentiment_task(sentence)

Output

[{'label': 'LABEL_0', 'score': 0.7535950541496277}]

Author

LeeMinTuan - BI team