RickyIG/legal-qwen25-3b-sft-exp10

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:3.1BQuant:BF16Ctx Length:32kPublished:May 20, 2026License:apache-2.0Architecture:Transformer Open Weights Warm

RickyIG/legal-qwen25-3b-sft-exp10 is a 3.1 billion parameter Qwen2.5-Instruct model, fine-tuned by RickyIG, leveraging Unsloth for accelerated training. This model is specifically optimized for legal applications, building upon the Qwen2.5 architecture with a 32768 token context length. Its primary differentiation lies in its specialized fine-tuning for legal tasks, making it suitable for processing and generating legal-specific text.

Loading preview...

RickyIG/legal-qwen25-3b-sft-exp10 Overview

This model is a specialized fine-tuned version of the Qwen2.5-3B-Instruct architecture, developed by RickyIG. It utilizes the unsloth/Qwen2.5-3B-Instruct-bnb-4bit as its base and was trained using Unsloth and Huggingface's TRL library, which enabled a 2x faster fine-tuning process. With 3.1 billion parameters and a 32768 token context length, it is designed for efficient performance.

Key Capabilities

  • Legal Domain Specialization: Fine-tuned specifically for legal applications, suggesting enhanced performance on legal texts and queries.
  • Efficient Training: Benefits from Unsloth's optimization, allowing for faster fine-tuning compared to standard methods.
  • Qwen2.5 Architecture: Inherits the robust capabilities of the Qwen2.5-Instruct base model.

Good For

  • Legal Text Processing: Ideal for tasks involving legal documents, research, and analysis.
  • Domain-Specific Applications: Suitable for developers building applications that require a strong understanding of legal language and concepts.
  • Resource-Efficient Deployment: Its 3.1 billion parameter size makes it a good candidate for deployment in environments where computational resources are a consideration, while still offering specialized capabilities.