Microsoft Supercomputer: This supercomputer has been built for training ChatGPT and Bing AI chatbots. Several thousand Nvidia A100 graphics chipsets have been supported in this.
AI Supercomputer: New developments on Artificial Intelligence are being seen all over the world. giant tech company Microsoft also did Open AI A supercomputer has been built together with It will be used to train Artificial Intelligence (AI) based chatbots. In the latest supercomputer, the company has several thousand Nvidia A100 Graphics chipset has been supported, so that ChatGPT and Bing AI can be trained.
In 2019, Microsoft has invested 1 billion dollars (about 8 thousand crore rupees) in OpenAI. Apart from this, the company also agreed to build a big and modern supercomputer. This supercomputer will be of great use in improving large language models like ChatGPT.
read this also- You can also use ChatGPT 4, just have to do this work
More accuracy with more capacity
The purpose of building supercomputers is to provide computing power for training large sets of AI models with data over a long period of time. “What we’ve learned from research is that the bigger the model, the more data you have and the longer you train,” said Nidhi Chappell, Microsoft’s product head for Azure high-performance computing and AI. can, the better the accuracy of the model.”
Required for AI chatbot
He further said that a strong support was needed to train big models (AI Chatbots) for a long time. Which means not only do you need the largest infrastructure, but you also need to be able to run it stably over a long period of time. Let us tell you that during the Build Developers Conference in 2020, Microsoft had said that it is making a supercomputer.
Powerful cloud setup is also needed
The company had informed that the supercomputer is being specially prepared for OpenAI. The supercomputer designed with Azure will play a key role in training the AI model. This will further encourage research on technology like ChatGPT. More than a supercomputer, a powerful cloud setup is also needed to train the AI model. Microsoft is already working on this.
