Qwen/Qwen1.5-72B-Chat
TEXT GENERATIONConcurrency Cost:4Model Size:72.3BQuant:FP8Ctx Length:32kPublished:Jan 30, 2024License:otherArchitecture:Transformer0.2K Cold

Qwen/Qwen1.5-72B-Chat is a 72.3 billion parameter, transformer-based decoder-only language model developed by Qwen. This chat-optimized model is part of the Qwen1.5 series, offering significant performance improvements in human preference and stable multilingual support with a 32K context length. It is designed for conversational AI applications requiring large-scale language understanding and generation.

Loading preview...

Qwen1.5-72B-Chat Overview

Qwen1.5-72B-Chat is a 72.3 billion parameter, transformer-based decoder-only language model from the Qwen1.5 series, serving as a beta version for Qwen2. This model builds upon previous Qwen iterations with several key enhancements, including improved human preference in chat models and robust multilingual capabilities for both base and chat variants. It supports a stable 32K context length across all model sizes, making it suitable for processing longer inputs and maintaining conversational coherence.

Key Capabilities

  • Enhanced Chat Performance: Demonstrates significant improvements in human preference for conversational tasks.
  • Multilingual Support: Offers native multilingual capabilities for both base and chat models.
  • Extended Context Window: Provides stable support for a 32K token context length, facilitating more complex and lengthy interactions.
  • Simplified Integration: No longer requires trust_remote_code for Hugging Face Transformers integration, streamlining deployment.
  • Architectural Foundation: Utilizes a Transformer architecture with SwiGLU activation, attention QKV bias, and group query attention, optimized for diverse language tasks.

Good For

  • Conversational AI: Ideal for building advanced chatbots and virtual assistants due to its chat optimization and human preference improvements.
  • Multilingual Applications: Suitable for use cases requiring understanding and generation in multiple languages.
  • Long-Context Tasks: Effective for applications that benefit from processing extensive textual information, such as document summarization or detailed question answering.
Finetunes
6 models