| Property | Value |
|---|---|
| Parameter Count | 12.2B |
| License | CC-BY-4.0 |
| Paper | Model Stock Methodology |
| Base Model | Mistral-Nemo-Instruct-2407 |
What is MN-Violet-Lotus-12B-GGUF?
MN-Violet-Lotus-12B-GGUF is a quantized version of the original MN-Violet-Lotus-12B model, specifically designed for creative writing and roleplaying applications. This model stands out for achieving an impressive 80.00 score on emotional intelligence benchmarks with 100% parsing at 8-bit quantization, making it particularly effective for character-driven narrative generation.
Implementation Details
The model is implemented using the Model Stock merge methodology, combining several specialized models including Violet_Twilight-v0.2, Lumimaid-v0.2-12B, Mahou-1.5-mistral-nemo-12B, and MN-12B-Lyra-v4. The merge was performed using mergekit, with Mistral-Nemo-Instruct-2407 serving as the base model anchor point.
- Utilizes SLERP blending for core roleplaying capabilities
- Implements Model Stock merge methodology for optimal weight distribution
- Features bfloat16 dtype with normalized parameters
Core Capabilities
- Enhanced creative writing and storytelling
- High emotional intelligence in character interactions
- Consistent personality adherence in roleplaying
- Balanced output length and coherence
Frequently Asked Questions
Q: What makes this model unique?
The model's distinctive feature is its combination of high emotional intelligence scores with creative writing capabilities, achieved through a carefully crafted merge of specialized models and quantization for practical deployment.
Q: What are the recommended use cases?
This model is particularly well-suited for creative writing, text adventures, roleplaying scenarios, and any applications requiring nuanced character interactions with consistent personality traits.