finbert-financial-news-lora

FinBERT Financial News Sentiment — LoRA Adapter

This repository contains a LoRA (PEFT) adapter fine-tuned for financial news sentiment analysis using the Sentiment Analysis for Financial News (FinancialPhraseBank) dataset.

Labels

• 0: negative
• 1: neutral
• 2: positive

Base model

• ProsusAI/finbert

Usage

import torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification
from peft import PeftModel

base_id = "ProsusAI/finbert"
adapter_id = "prithvi1029/finbert-financial-news-lora"

tok = AutoTokenizer.from_pretrained(base_id)
base = AutoModelForSequenceClassification.from_pretrained(
    base_id, num_labels=3
)
model = PeftModel.from_pretrained(base, adapter_id)

text = "Company shares surged after strong quarterly earnings."
inputs = tok(text, return_tensors='pt', truncation=True)

with torch.no_grad():
    logits = model(**inputs).logits
    pred = logits.argmax(-1).item()

print('Prediction:', pred)

Disclaimer

This model is for research and educational purposes only and not financial advice.

Training data

• prithvi1029/sentiment-analysis-for-financial-news