Lura-1.0

Introducing Lura-1.0, a lightweight 500M parameter model fine-tuned on top of the Qwen 2.5 0.5B base architecture.

By leveraging a high-quality, custom-curated dataset, Lura-1.0 achieves a massive jump in performance over standard base models of its size, offering impressive efficiency in a compact footprint.

Highlights

• Targeted Performance: Excels at basic mathematical reasoning, clean code generation, and everyday assistant tasks.
• Custom Fine-Tuning: Trained on dense, high-quality instruction pairs for optimal parameter efficiency.
• Ultra-Lightweight & Local: With just 500M parameters, it runs blazingly fast locally on standard hardware and edge devices with minimal RAM.

Model Specs

• Base Architecture: Qwen 2.5 (0.5B)
• Parameter Count: 500M
• Use Cases: Simple code, math problems, low-latency local assistants

By SEN-AGI