Hugging Face Evathene-v1.1 by sophosympatheia is a 72.7 billion parameter merged language model, combining Nexusflow/Athene-V2-Chat and EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2. This model is specifically designed and optimized for roleplaying and storytelling tasks, offering competent writing capabilities. It is an uncensored model, intended for users seeking a formal writing style in narrative generation.
Evathene-v1.1: A 72.7B Parameter Model for Roleplaying and Storytelling
Evathene-v1.1, developed by sophosympatheia, is a 72.7 billion parameter language model created by merging Nexusflow/Athene-V2-Chat with EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2 using the TIES merge method. This version is an update to the original Evathene release, benefiting from the EVA-Qwen2.5-72B-v0.2 component to enhance its capabilities.
Key Capabilities & Characteristics
- Optimized for Narrative Generation: Specifically designed for roleplaying and storytelling, providing competent and intelligent prose.
- Uncensored Output: The model is uncensored, allowing for a wide range of content generation without built-in restrictions.
- Formal Writing Style: Users may find this version particularly suitable for tasks requiring a more formal writing style compared to other Evathene iterations.
- Merge Composition: Built upon the Qwen/Qwen2.5-72B base, integrating strengths from both Athene-V2-Chat and EVA-Qwen2.5-72B-v0.2.
Usage Recommendations
Evathene-v1.1 is recommended for developers and users focused on generating detailed and consistent narratives for roleplaying scenarios or creative writing. It is noted for its ability to respond competently to out-of-character (OOC) prompts and rewrite requests, suggesting good adherence to user guidance. The model's licensing terms are complex due to its merged nature, advising users to consult legal counsel for commercial applications.