Alibaba Cloud announced the full release of the Ali large model family, and the official upgrade and release of Tongyi Qianwen 2.0, which increased the number of large model parameters to 100 billion yuan, reaching the highest level in the world.
The performance exceeds GPT-3.5 and is accelerating to catch up with GPT-4.
Tongyi Qianwen 2.0 has been greatly improved in terms of reading comprehension, logical thinking, data, etc. And this time it supports features such as voice conversations.
At today's conference, Ali showed application models such as Tongyi Stardust, Tongyi Lingcode, Tongyi Xiaomi, Tongyi Dianjin, Tongyi Zhiwen, Tongyi Farui, and Tongyi Renxin.
Ali will open source the Qwen72B 72 billion parameter model of Tongyi Qianwen in November, and 72B will become the largest open-source large model in China. Now the cumulative number of downloads of the model has exceeded 1 million!
In the future, all software can be connected to large models, and then upgraded, Tongyi Qianwen has open capabilities, and can create its own exclusive industry GPT for each enterprise.
Over the years, Alibaba Cloud has built a complete AI infrastructure and a complete AI large model training platform, which can provide you with a very complete model inference service.
ongyi Lingma Al Assistant for Coding
Based on the Tongyi large model, it has been trained on massive excellent open source code data and R&D knowledge, providing intelligent code generation, R&D intelligent Q&A capabilities, and supporting mainstream IDE line-level and function-level continuation. Unit Test Generation Code Comment Generation Debugging Troubleshooting R&D Intelligent Q&A
The world is changing, a new era of AI is coming, you see it or not, the power of large language models is there, the key is how to use it. If you had a "Thousand Questions of the Senses", what would you do with it? Welcome to leave a message in the comment area.