deberta-v3-base-absa-v1.1
| Property | Value |
|---|---|
| Parameter Count | 184M |
| License | MIT |
| Author | yangheng |
| Paper | Research Paper |
| Datasets | 8 (including Laptop14, Restaurant14, MAMS, etc.) |
What is deberta-v3-base-absa-v1.1?
This is a specialized DeBERTa-v3 model fine-tuned for aspect-based sentiment analysis (ABSA). Built on Microsoft's DeBERTa-v3-base architecture, it's been trained on over 180,000 examples across 8 diverse datasets, making it particularly robust for analyzing sentiment in context of specific aspects of products or services.
Implementation Details
The model is implemented using the FAST-LCF-BERT architecture and is powered by PyABSA, an open-source tool for aspect-based sentiment analysis. It processes text using a [CLS] and [SEP] token structure to identify and analyze sentiment towards specific aspects within text.
- Based on microsoft/deberta-v3-base architecture
- Trained on 30k+ ABSA samples + augmented data
- Supports multiple domains including restaurants, laptops, and retail
- Implements efficient text pair classification
Core Capabilities
- Aspect-specific sentiment classification
- Multi-domain sentiment analysis
- Fine-grained opinion mining
- Support for both training and inference
Frequently Asked Questions
Q: What makes this model unique?
This model's uniqueness lies in its comprehensive training across 8 different datasets and its specialized ability to perform aspect-based sentiment analysis, making it particularly effective for detailed sentiment analysis tasks where context matters.
Q: What are the recommended use cases?
The model is ideal for analyzing customer reviews, product feedback, and service evaluations where understanding sentiment about specific aspects is crucial. It's particularly well-suited for applications in retail, hospitality, and product analysis.
