SlerpE/NightyGurps-14b-v1.1

Hugging Face
TEXT GENERATIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:14.8BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 21, 2024License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Warm

SlerpE/NightyGurps-14b-v1.1 is a 14.8 billion parameter, monolingual Russian language model fine-tuned from Qwen 2.5 14b. It is specifically designed for GURPS tabletop role-playing game assistance, acting as a consultant, GM assistant, or even a solo GM. This model excels at providing GURPS rule-based guidance and facilitating role-play sessions within the GURPS system.

Loading preview...

NightyGurps-14b-v1.1: GURPS Role-Playing Assistant (Russian Monolingual)

NightyGurps-14b-v1.1 is a specialized 14.8 billion parameter language model developed by SlerpE, fine-tuned from the Qwen 2.5 14b architecture. Its core purpose is to serve as a dedicated assistant for the GURPS (Generic Universal RolePlaying System) tabletop RPG, operating exclusively in Russian.

Key Capabilities

  • GURPS Rule Consultation: Provides detailed advice and clarification on GURPS rules, acting as a consultant or assistant to the Game Master.
  • Game Master (GM) Assistance: Supports GMs in running GURPS campaigns.
  • Solo Play GM: Capable of acting as the Game Master for solo GURPS sessions.
  • Optimized for Role-Play: Specifically trained on an augmented dataset derived from the GURPS Basic Set rulebook to enhance its role-playing capabilities.

Usage Recommendations

  • For consultant/assistant mode, a temperature setting of 0.3-0.5 is recommended for factual accuracy.
  • For Game Master mode, a higher temperature of 0.7-1.2 is suggested to encourage creative storytelling and dynamic interactions.
  • The model utilizes the ChatML format for prompting, requiring specific system, user, and assistant tags.

Future Development

SlerpE plans to expand the training dataset significantly and release smaller (NightyGurps-Nano, based on Qwen 2.5 3b) and larger (NightyGurps-Large, based on Qwen 2.5 32b) versions to cater to different performance and resource requirements.