Undi95/Llama-3-LewdPlay-8B
Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kPublished:Apr 23, 2024License:cc-by-nc-4.0Architecture:Transformer0.0K Open Weights Warm

Undi95/Llama-3-LewdPlay-8B is an 8 billion parameter Llama 3 based model fine-tuned by Undi95. It is specifically optimized for roleplay (RP) tasks, leveraging multiple specialized RP datasets. This model is designed to generate engaging and contextually relevant responses for character-driven interactions.

Loading preview...

Overview

Undi95/Llama-3-LewdPlay-8B is an 8 billion parameter model built upon the meta-llama/Meta-Llama-3-8B-Instruct architecture. Developed by Undi95, this model has been extensively fine-tuned on a diverse collection of roleplay (RP) datasets to enhance its capabilities in generating interactive and character-driven narratives.

Key Capabilities

  • Specialized Roleplay Generation: Fine-tuned on multiple RP datasets, including Undi95/toxic-dpo-v0.1, NobodyExistsOnTheInternet/ToxicQAFinal, Aesir, LimaRP, and cleaned versions of Pippa and Bluemoon, making it highly proficient in roleplaying scenarios.
  • Llama 3 Prompt Format: Utilizes the standard Llama 3 prompt template, ensuring compatibility with existing Llama 3 workflows and tools.
  • Optimized for Character Interaction: Designed to perform best with well-written character cards, including example messages and memory, to facilitate rich and consistent roleplay.

Good For

  • Roleplaying Applications: Ideal for applications requiring detailed and engaging character interactions.
  • Interactive Storytelling: Suitable for generating dynamic narratives where user input drives the story.
  • Creative Content Generation: Can be used for creating dialogue and scenarios for virtual characters or chatbots.
Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p