阿里云近日宣布开源通义千问的三个模型，包括720亿参数模型Qwen-72B、18亿参数模型Qwen-1.8B以及音频大模型Qwen-Audio。这些模型已经在阿里魔搭社区上线，为开发者们提供了丰富的资源。
据了解，通义千问是阿里云推出的一款人工智能语言模型，具有强大的自然语言处理能力。其中，720亿参数模型Qwen-72B和18亿参数模型Qwen-1.8B是目前市面上最大的中文预训练模型之一，而音频大模型Qwen-Audio则是一款专门针对音频数据进行处理的模型。
除了预训练模型外，阿里云还同步推出了对应的对话模型。这些对话模型面向72B、1.8B两种规模的模型提供了4bit/8bit量化版模型，便于开发者们进行推理训练。这一举措将极大地降低开发者们的使用门槛，提高开发效率。
阿里云表示，开源这些模型的目的是希望通过共享资源，推动人工智能技术的发展。未来，阿里云将继续加大对人工智能领域的投入，推出更多优秀的产品和服务。
英语如下：
Title: Aliyun Open Sources Tongyi Qianwen Model to Boost AI Developer Innovation
Keywords: Aliyun open source, Tongyi Qianwen model, dialogue model
Content: Aliyun recently announced the open-sourcing of three models from the Tongyi Qianwen series, including the 72 billion parameter model Qwen-72B, the 1.8 billion parameter model Qwen-1.8B, and the audio model Qwen-Audio. These models are now available on the Aliyun Magic Build Community, providing developers with a wealth of resources.
Tongyi Qianwen is an artificial intelligence language model launched by Aliyun, boasting strong natural language processing capabilities. The 72 billion parameter model Qwen-72B and the 1.8 billion parameter model Qwen-1.8B are currently among the largest pre-trained Chinese language models available, while the audio model Qwen-Audio is specifically designed for processing audio data.
In addition to the pre-trained models, Aliyun has also released corresponding dialogue models. These dialogue models provide 4bit/8bit quantized versions for both the 72B and 1.8B scale models, making it easier for developers to conduct inference training. This move will greatly reduce the barriers to entry for developers and improve development efficiency.
Aliyun stated that the purpose of open-sourcing these models is to promote the development of artificial intelligence technology through resource sharing. In the future, Aliyun will continue to increase investment in the field of artificial intelligence and launch more excellent products and services.
【来源】https://www.modelscope.cn/headlines/article/269
Views: 1