FireBERT – Indian Financial News Classifier

FireBERT is a fine-tuned version of ProsusAI/finbert, designed specifically for the Indian financial ecosystem. It classifies news headlines into one of three categories:

Macro – Government policies, RBI updates, GDP, inflation, and broader economic signals
Stock – Company-specific updates like earnings, partnerships, and expansion plans
Market – Index moves, sectoral shifts, and overall investor sentiment

This model was developed to support analysts, investors, and researchers in quickly understanding the category of any Indian financial news headline.

Model Details

Developed by: Yogeeth GK (Fire)
Base Model: ProsusAI/finbert
Finetuned on: Indian financial headlines (custom dataset)
Language(s): English
License: Apache 2.0

Update Policy

FireBERT is updated and improved on the 24th of every month with better training data, cleaner labels, and domain-specific refinements. This ensures relevance with fast-evolving financial news in India.

About Me

I’m Yogeeth GK, a student at NIT Durgapur, passionate about using AI and VLSI to build tools to serve.FireBERT reflects my aspiration to empower the world with clarity and intelligence in finance.

How to Use

from transformers import AutoTokenizer, AutoModelForSequenceClassification

model_id = "yoganfire/firebert"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForSequenceClassification.from_pretrained(model_id)

def predict(text):
    inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True)
    outputs = model(**inputs)
    label = outputs.logits.argmax(dim=1).item()
    return ["Macro", "Stock", "Market"][label]

Model Card Contact

Author: Yogeeth GK
Affiliation: NIT Durgapur
Website: https://stockinfoai.netlify.app/

For full documentation and monthly release notes, visit the model repository.