BERT Stock Sentiment Classifier (Fine-Tuned)

This is a BERT-based model fine-tuned on a dataset of stock market news headlines to perform sentiment analysis. The labels are:

positive
neutral
negative

The model is intended for financial news and headlines, especially useful for trading, sentiment scoring, or market analysis pipelines.

🧾 How It Was Trained

Base model: bert-base-uncased
Dataset: Custom scraped Finviz news headlines
Labels: Generated using FinBERT, mapped to positive, neutral, negative
Training: 3 epochs, batch size 16, learning rate 2e-5

🛠 Usage

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

tokenizer = AutoTokenizer.from_pretrained("hasnain43/bert-stock-sentiment-v1")
model = AutoModelForSequenceClassification.from_pretrained("hasnain43/bert-stock-sentiment-v1")
model.eval()

label_map = {0: "negative", 1: "neutral", 2: "positive"}

def predict_sentiment(text):
    inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True)
    with torch.no_grad():
        outputs = model(**inputs)
        logits = outputs.logits
        prediction = torch.argmax(logits, dim=1).item()
    return label_map[prediction]

predict_sentiment("Tesla stock drops after disappointing delivery numbers.")