flozi00/Llama-2-13B-german-assistant-v3

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

The flozi00/Llama-2-13B-german-assistant-v3 is a 13 billion parameter Llama 2-based model developed by flozi00, specifically fine-tuned for German instruction following and conversational tasks. It utilizes Open Assistant-style tokens and was trained on a deduplicated, cleaned dataset without code. This model excels in generating German responses for assistant-like interactions and general conversations.

Loading preview...

Model Overview

The flozi00/Llama-2-13B-german-assistant-v3 is a 13 billion parameter language model built upon the Llama 2 architecture. Developed by flozi00, this version is specifically fine-tuned for German language tasks, focusing on instruction following and conversational capabilities.

Key Characteristics

  • German Language Focus: Optimized for generating responses in German, making it suitable for German-centric applications.
  • Instruction Following: Designed to accurately follow instructions provided in German.
  • Conversational Abilities: Excels in engaging in natural, assistant-style conversations.
  • Training Data: Fine-tuned on a carefully deduplicated and cleaned dataset, ensuring high-quality German text without embedded code.
  • Tokenization: Utilizes Open Assistant-style tokens (<|prompter|>, <|endoftext|>, <|assistant|>) for structured interactions.
  • Sustainable Training: The model was trained using hardware powered by 100% renewable energy.

Use Cases

This model is particularly well-suited for:

  • German-speaking AI assistants: Providing conversational support and answering queries in German.
  • Instruction-based tasks: Executing commands and generating content based on German instructions.
  • Chatbots: Developing interactive German chatbots for various applications.

Note: The developer recommends using V4 of this model for improved performance.