openSUSE/Cavil-Qwen3-4B is a LoRA fine-tune of the Qwen3-4B model, developed by openSUSE for specialized legal text classification. This model is specifically adapted for use with Cavil, focusing on analyzing and categorizing legal documents. Its primary differentiator is its fine-tuning on the cavil-legal-text dataset, making it highly effective for legal domain-specific tasks.
Loading preview...
Overview
openSUSE/Cavil-Qwen3-4B is a specialized language model created by openSUSE. It is a LoRA (Low-Rank Adaptation) fine-tune of the base Qwen3-4B model, specifically designed for legal text classification tasks within the Cavil framework.
Key Capabilities
- Legal Text Classification: The model is fine-tuned on the
cavil-legal-textdataset, making it proficient in classifying and understanding legal documents. - Cavil Integration: Developed for seamless integration and use with the Cavil project.
Training Details
The fine-tuning process is detailed in a blog post. The LoRA adapter was trained exclusively on the cavil-legal-text dataset, which was curated by SUSE LLC.
Licensing
The base Qwen3-4B model operates under the Apache-2.0 license. The Cavil LoRA adapter, including its training data, is licensed by SUSE LLC under Apache-2.0, with the dataset itself being under GPL-2.0-or-later.
Example Implementation
An example implementation demonstrating how to deploy and use this model is available at llm-lawyer.