TheBloke/Pygmalion-13B-SuperHOT-8K-fp16

TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kLicense:otherArchitecture:Transformer0.0K Cold

TheBloke/Pygmalion-13B-SuperHOT-8K-fp16 is a 13 billion parameter conversational language model, a merge by TehVenom of PygmalionAI's Pygmalion 13B with Kaio Ken's SuperHOT 8K. This model is fine-tuned for fictional conversation and roleplay, leveraging an extended context length of 8192 tokens. It is designed for GPU inference in fp16 PyTorch format, excelling in dialogue generation within a specified character persona.

Loading preview...

Model Overview

This model, Pygmalion-13B-SuperHOT-8K-fp16, is a 13 billion parameter conversational language model. It is a merge by TehVenom, combining PygmalionAI's Pygmalion 13B with Kaio Ken's SuperHOT 8K. The model is provided in fp16 PyTorch format, suitable for GPU inference.

Key Capabilities

  • Extended Context Window: Achieves an 8192-token context length, enabled by the SuperHOT 8K merge and trust_remote_code=True during inference.
  • Conversational AI: Fine-tuned for dialogue, specifically designed for fictional conversations and character roleplay.
  • Persona-driven Responses: Trained to generate responses based on a provided character persona and chat history, using a specific prompting format.

Intended Use Cases

  • Fictional Conversation: Ideal for entertainment purposes, such as creating interactive story experiences or character-driven chatbots.
  • Roleplay Scenarios: Excels at maintaining a consistent character persona throughout a dialogue.

Limitations

  • Not Safety-Tuned: The model was not fine-tuned for safety or harmlessness and may produce socially unacceptable or factually incorrect text.
  • Specific Prompting Format: Requires a particular [CHARACTER]'s Persona:, <START>, [DIALOGUE HISTORY], You:, [CHARACTER]: format for optimal performance.