| Property | Value |
|---|---|
| Parameter Count | 184M |
| Model Type | DeBERTa-v2 |
| License | MIT |
| Datasets | 4 Financial datasets |
What is Finance-Sentiment-Classification?
Finance-Sentiment-Classification is a specialized DeBERTa-based model designed for sentiment analysis in financial contexts. Initially trained on over 1 million Amazon reviews, this model has been fine-tuned on four distinct financial datasets to provide accurate sentiment classification for financial text. The model categorizes text into three sentiment classes: negative, neutral, and positive.
Implementation Details
The model leverages the DeBERTa-v2 architecture and is implemented using PyTorch and the Transformers library. It has been trained on multiple financial datasets including financial_phrasebank, kaggle-financial-sentiment, twitter-financial-news-sentiment, and auditor_sentiment, ensuring robust performance across various financial contexts.
- Utilizes the DeBERTa-v2 architecture
- Implements three-way classification (negative, neutral, positive)
- Supports batch processing with padding
- Maximum sequence length of 65 tokens
Core Capabilities
- Financial text sentiment analysis
- Real-time sentiment prediction
- Probability distribution for sentiment classes
- Support for both GPU and CPU inference
Frequently Asked Questions
Q: What makes this model unique?
This model combines the power of DeBERTa-v2 with specialized training on financial datasets, making it particularly effective for financial sentiment analysis. The initial training on Amazon reviews followed by fine-tuning on financial data provides robust sentiment detection capabilities.
Q: What are the recommended use cases?
The model is ideal for analyzing financial news, market reports, social media posts about stocks, and auditor statements. It can be integrated into trading systems, market analysis tools, or financial monitoring applications.