wenge-research/yayi-7b-llama2

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Jul 21, 2023Architecture:Transformer0.0K Cold

The YaYi-7B-Llama2 model by wenge-research is a 7 billion parameter Llama 2-based language model optimized for Chinese multi-domain tasks. It was fine-tuned on high-quality domain-specific data covering media publicity, public opinion analysis, public safety, financial risk control, and urban governance. This model enhances foundational Chinese language capabilities and domain analysis, supporting multi-turn conversations and plugin integration.

Loading preview...

YaYi-7B-Llama2: Chinese Optimized Llama 2 Model

The YaYi-7B-Llama2 is a 7 billion parameter language model developed by Beijing Wenge Technology Co.,Ltd., built upon the Llama 2 architecture. It has been specifically optimized for Chinese multi-domain tasks through extensive instruction fine-tuning.

Key Capabilities & Features

  • Domain-Specific Expertise: Fine-tuned on millions of high-quality, artificially constructed domain data points across five key areas:
    • Media publicity
    • Public opinion analysis
    • Public safety
    • Financial risk control
    • Urban governance
  • Enhanced Chinese Language Processing: Steadily improved foundational Chinese language capabilities and domain analysis during its iterative development.
  • Multi-turn Conversation: Supports multi-turn conversational interactions.
  • Plugin Integration: Includes partial plugin capabilities.
  • User Feedback Optimization: Performance and security were refined through continuous manual feedback from hundreds of internal users.

When to Use This Model

This model is particularly well-suited for applications requiring strong Chinese language understanding and generation within specific domains such as public sector analysis, financial services, and urban management. Its Llama 2 base combined with targeted Chinese optimization makes it a robust choice for developers focusing on these areas.