Model Overview

Lunzima/NQLSG-Qwen2.5-14B-MegaFusion-v9.2 is a 14.8 billion parameter language model developed by Lunzima. It was created using the advanced Model Stock merge method, which combines multiple pre-trained models to synthesize their capabilities into a single, more robust model. This particular iteration builds upon the foundation of Lunzima/NQLSG-Qwen2.5-14B-MegaFusion-v9.1 and its Alpaca-LoRA fine-tuned variant.

Merge Details

This model is a sophisticated fusion of several distinct language models, aiming to consolidate their respective strengths. The merge process incorporated:

• Cran-May/tempmotacilla-cinerea-0308
• suayptalha/Lix-14B-v0.1
• suayptalha/Lamarckvergence-14B
• Lunzima/NQLSG-Qwen2.5-14B-MegaFusion-v9.1

The merging was performed using mergekit, a tool for combining language models, with a specific YAML configuration that utilized bfloat16 precision and an int8_mask for optimization. The tokenizer source was derived from the base model, ensuring consistent tokenization across the merged architecture.

Key Characteristics

This model is designed for general-purpose language tasks, benefiting from the diverse training data and architectural nuances of its merged components. Its 14.8 billion parameters and 32768 token context window make it suitable for handling complex prompts and generating extensive responses.