Trappu/Nemo-Picaro-12B
Nemo-Picaro-12B is a 12 billion parameter language model developed by Trappu, fine-tuned from royallab/MN-LooseCannon-12B-v2. This model is specifically trained on a unique dataset focused on storywriting and scenario prompting, making it a specialized tool for creative narrative generation. It is characterized by its "one-trick pony" nature, excelling in specific creative tasks despite acknowledged limitations like rampant impersonation and a lack of general intelligence. The model supports a context length of 32768 tokens.
Loading preview...
Overview
Trappu/Nemo-Picaro-12B is a 12 billion parameter language model developed by Trappu, fine-tuned from royallab/MN-LooseCannon-12B-v2. This model is uniquely specialized for creative storywriting and scenario prompting, trained on a custom dataset devoid of synthetic data. It is designed to generate narratives based on specific scenarios and tags, making it a niche tool for particular creative applications.
Key Characteristics
- Specialized Training: Focused exclusively on storywriting and scenario prompting (e.g.,
[ Scenario: bla bla bla; Tags: bla bla bla ]). - Context Length: Supports a substantial context window of 32768 tokens.
- Acknowledged Limitations: The developer notes its "one-trick pony" nature, with potential issues like rampant impersonation and a lack of general intelligence, making it challenging for average LLM users.
- Prompting Format: Utilizes ChatML format without system prompts, with specific templates provided for platforms like SillyTavern.
Good For
- Dedicated Story Generation: Ideal for users specifically seeking a model for creative narrative development and scenario-based writing.
- Experimental Creative Projects: Suitable for developers and enthusiasts who can manage its specific quirks and limitations to leverage its creative strengths.
- SillyTavern Users: Pre-configured instruct and context templates, along with recommended settings, are available for seamless integration with SillyTavern.