After Perplexity, Microsoft introduces DeepSeek R1 on Azure AI Foundry and GitHUB – Crypto News – Crypto News
Connect with us
After Perplexity, Microsoft introduces DeepSeek R1 on Azure AI Foundry and GitHUB After Perplexity, Microsoft introduces DeepSeek R1 on Azure AI Foundry and GitHUB

Technology

After Perplexity, Microsoft introduces DeepSeek R1 on Azure AI Foundry and GitHUB – Crypto News

Published

1 year ago

on

Tech giant Microsoft has added DeepSeek R1 to the model catalog on Azure AI Foundry and GitHub, expanding its collection of over 1,800 AI models.

In a blog post, Microsoft said that this new addition comes as part of the company’s ongoing efforts to provide a wide variety of AI solutions, including frontier, open-source, industry-specific, and task-oriented models.

“DeepSeek R1 is now accessible via the Azure AI Foundry platform, offering businesses a way to integrate advanced AI technologies in a scalable, enterprise-ready environment,” the blog post stated.

The model is designed to improve AI reasoning, providing developers with a cost-effective solution for using advanced AI capabilities without the need for extensive infrastructure. As per the company, DeepSeek R1 offers tools that allow for quick experimentation, benchmarking, and performance evaluation. These tools aim to streamline the development process, enabling faster iteration and integration of AI into applications.

Azure AI Foundry’s model catalog is positioned as a resource for companies to access a wide range of AI models in one place. The platform also includes built-in features such as model evaluation tools, making it easier for developers to test and scale their AI applications.

Microsoft highlighted that security remains a focus, and DeepSeek R1 has undergone various safety evaluations, including model behaviour assessments and security reviews. The platform incorporates safeguards like Azure AI Content Safety, which provides automatic content filtering, and a Safety Evaluation System, which allows customers to test their applications before deployment.

For developers looking to use DeepSeek R1, the process involves signing up for an Azure account and navigating the model catalog to find the model. After deployment, users are given an inference API and key, with access to a playground for testing the model. The company has also announced plans to offer distilled versions of DeepSeek R1 that can run locally on Copilot+ PCs, a feature expected to be available soon.

Related Topics:
Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Trending