ArliAI/Llama-3.1-8B-ArliAI-RPMax-v1.2
Llama-3.1-8B-ArliAI-RPMax-v1.2 by ArliAI is an 8 billion parameter language model, a variant of Meta's Llama-3.1, specifically fine-tuned for creative writing and roleplay. It is part of the RPMax series, trained on a diverse and deduplicated dataset to ensure high creativity and non-repetitive outputs. This model excels at generating varied character interactions and scenarios, making it suitable for immersive narrative applications.
Loading preview...
ArliAI/Llama-3.1-8B-ArliAI-RPMax-v1.2 Overview
This model is an 8 billion parameter variant of Meta's Llama-3.1, developed by ArliAI as part of their RPMax series. It is specifically fine-tuned for creative writing and roleplay (RP), focusing on generating highly creative and non-repetitive content.
Key Capabilities & Features
- Enhanced Creativity: Trained on a diverse and deduplicated dataset, ensuring the model avoids repetitive character traits or situations, leading to more dynamic and unique interactions.
- Roleplay Optimization: Designed to understand and appropriately act as various characters in diverse scenarios, providing a distinct and less "in-bred" feel compared to other RP models.
- Llama 3.1 Base: Leverages the strong foundational capabilities of the Llama 3.1 architecture.
- Training Details: Underwent 1 epoch of training on 2x3090Ti for approximately 1 day, utilizing a 64-rank 128-alpha LORA with a low learning rate of 0.00001 and gradient accumulation of 32 to minimize repetition sickness.
- Context Length: Supports a context length of 128K tokens.
Recommended Use Cases
- Interactive Storytelling: Ideal for applications requiring dynamic and varied narrative generation.
- Character Roleplay: Excels in scenarios where the model needs to embody distinct personalities and respond appropriately to user inputs.
- Creative Content Generation: Suitable for generating unique dialogues, plotlines, and descriptive text without falling into repetitive patterns.
Users can access the model and find more information at ArliAI.