This model is fine-tuned to perform aspect-based sentiment analysis (ABSA), identifying sentiment (positive, neutral, or negative) and aspects category such as food, service, and ambiance in restaurant reviews. It builds on the open-source Gemma model (9B parameters) with LoRA for parameter-efficient fine-tuning, making it lightweight and accessible without compromising performance.
-
Key Features
Aspect-Aware Sentiment Detection: Goes beyond overall sentiment analysis to detect sentiment for specific aspects category in text.
-
Intended Use
Customer Feedback Analysis: Gleaning actionable insights from reviews.
Business Intelligence: Identifying key areas for improvement based on customer sentiment.
NLP Research: As a benchmark for aspect-based sentiment analysis tasks.
This gemma2 model was trained 2x faster with
Unsloth and Huggingface's TRL library.
