ParasiticRogue/Merged-RP-Stew-V2-34B

TEXT GENERATIONConcurrency Cost:2Model Size:34BQuant:FP8Ctx Length:32kPublished:Mar 29, 2024License:apache-2.0Architecture:Transformer0.1K Open Weights Cold

ParasiticRogue/Merged-RP-Stew-V2-34B is a 34 billion parameter language model created by ParasiticRogue using mergekit, combining several pre-trained models including Nous-Capybara, Tess-34B, Nontoxic-Bagel, PiVoT-SUS-RP, NyakuraV2, and CausalLM-RP. This merge is specifically designed for enhanced character and lore understanding, diverse chat and roleplay, and engaging storytelling. It leverages a 32768 token context length, making it suitable for complex, long-form interactive narratives.

Loading preview...

Overview

ParasiticRogue/Merged-RP-Stew-V2-34B is a 34 billion parameter model created by ParasiticRogue through a sophisticated merge of multiple specialized language models using mergekit. The merge prioritizes specific strengths from its constituent models, such as Nous-Capybara for general aptitude and context preservation, Tess-1.5 for character and lore understanding, and Nontoxic-Bagel, PiVoT-SUS-RP, Nyakura, and CausalLM-RP for diverse chat, roleplay, and storytelling capabilities.

Key Capabilities

  • Enhanced Roleplay & Storytelling: Optimized for engaging and diverse interactive narratives.
  • Character & Lore Understanding: Improved ability to grasp and maintain character personas and intricate lore details.
  • Long Context Handling: Preserves longer context lengths, beneficial for extended conversations and complex scenarios.
  • Customizable System Prompts: Designed to respond effectively to detailed system prompts for fine-grained control over AI behavior.

Recommended Usage

This model is particularly well-suited for applications requiring rich, interactive, and consistent character-driven interactions. It includes specific chat commandments and prompt formatting (Chat-Vicuna) to guide the AI's responses, encouraging vivid detail, logical consistency, and adherence to persona. Specialized EXL2 and GGUF versions are also available for various deployment needs.