Local-Novel-LLM-project/kagemusya-7B-v1
kagemusya-7B-v1 is a 7 billion parameter language model developed by Local-Novel-LLM-project, based on the Ninja-V2 architecture. This model is specifically fine-tuned for novel generation, having been trained on approximately 1GB of web novel data. While its general chat capabilities are reduced, it excels at continuing fictional narratives and generating story content. It supports a context length of 4096 tokens, making it suitable for creative writing tasks.
Loading preview...
Model Overview
kagemusya-7B-v1 is a 7 billion parameter language model developed by Local-Novel-LLM-project, built upon the Ninja-V2 base model. Its primary distinction lies in its specialized training for novel generation, utilizing approximately 1GB of web novel data. This focused training has optimized its ability to create and continue fictional narratives.
Key Capabilities
- Novel Generation: Excels at continuing stories and generating new narrative content based on provided prompts.
- Contextual Storytelling: Demonstrates the ability to maintain narrative flow and context within a story.
- Genre Adaptation: Can be guided towards specific genres (e.g., school life, isekai, harem) by including keywords like "あらすじ:" (synopsis) and "ジャンル:" (genre) at the beginning of the input.
Limitations
- Reduced Chat Ability: The model's general chat capabilities are significantly diminished due to its specialization in novel generation.
Example Usage
Users can provide an initial story prompt, and the model will generate a continuation. For instance, providing "メロスは激怒した。必ず、かの邪智暴虐の王を除かなければならぬと決意した。メロスには、" (Melos was enraged. He resolved that he must eliminate the tyrannical king. Melos has the...) will result in a narrative continuation. While its primary strength is creative writing, it can also provide concise answers to direct questions, though this is not its optimized function.