DialoGPT-Financial-Market-Sentiment-Trading-Assistant
Fine-tuned DialoGPT-small for financial news sentiment analysis and market sentiment interpretation for trading and investment decisions.
Overview
- Base Model: microsoft/DialoGPT-small (117M parameters)
- Fine-tuning Method: LoRA (4-bit quantization)
- Dataset: Financial news sentiment dataset (1.5K samples)
- Training: 3 epochs with optimized hyperparameters
Key Features
- Real-time financial news sentiment analysis
- Market sentiment interpretation for trading decisions
- Bullish/Bearish/Neutral sentiment classification
- Conversational interface for market analysis
- Optimized for trading desks and investment research
Usage
from transformers import AutoTokenizer, AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("sweatSmile/DialoGPT-Financial-Market-Sentiment-Trading-Assistant")
tokenizer = AutoTokenizer.from_pretrained("sweatSmile/DialoGPT-Financial-Market-Sentiment-Trading-Assistant")
# Market sentiment analysis example
prompt = "<|user|> What's the market sentiment for this news: Tesla stock surges after record delivery numbers <|bot|>"
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=150, pad_token_id=tokenizer.eos_token_id)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
Applications
- Trading desk sentiment analysis
- Algorithmic trading signal generation
- Investment research and market analysis
- Hedge fund market sentiment monitoring
- Financial news interpretation for portfolio decisions
Training Details
- LoRA rank: 8, alpha: 16
- 4-bit NF4 quantization with fp16 precision
- Learning rate: 3e-4 with linear scheduling
- Batch size: 8, Max length: 256 tokens
- 3 epochs on curated financial sentiment data
Specialized for high-frequency market sentiment analysis in trading environments.