bluejude10/Smoothie-Qwen3-8B-KR-Self-Driving-Legal-v3
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:32kPublished:Feb 21, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

Smoothie-Qwen3-8B-KR-Self-Driving-Legal-v3 by bluejude10 is a QLoRA fine-tuned 8B parameter Qwen3-based model, specifically optimized for Korean self-driving legal domain Q&A. This experimental model was developed to test if domain-specific fine-tuning could enable a smaller model to outperform larger base models in a RAG pipeline. It focuses on legal interpretation within the autonomous driving sector, but testing revealed that fine-tuning in this context led to lower accuracy compared to non-fine-tuned models.

Loading preview...