Qwen: Alibaba Cloud's multilingual language models

Qwen is a family of transformer-based large language models ranging from 1.8B to 72B parameters, developed by Alibaba Cloud. The models are pretrained on up to 3 trillion tokens of multilingual data with focus on Chinese and English content across various domains. Each model variant includes both base versions for general language modeling and chat versions fine-tuned for conversational interactions using supervised fine-tuning and reinforcement learning from human feedback. The models support features like tool usage, code interpretation, and mathematical problem solving, with quantized versions available in Int4 and Int8 formats for reduced memory requirements.