TL;DR:
- Dell partners with Meta (Facebook’s parent company) to facilitate the on-premises deployment of Llama 2 AI.
- Dell aims to become the preferred provider for businesses seeking on-premises AI solutions.
- The collaboration centers around Dell’s Validated Design for Generative AI portfolio.
- Dell offers pre-tested hardware builds, co-engineered with Nvidia, along with deployment and configuration guidance.
- Dell integrates Llama 2 models into its system sizing tools for tailored configurations.
- Jeff Boudreau, Dell’s Chief AI Officer, emphasizes the transformative potential of generative AI models.
- Llama 2 comes in three sizes (7 billion, 13 billion, and 70 billion parameters) with varying hardware requirements.
- Llama 2 is available for research and limited commercial use.
- Meta has previously partnered with Microsoft and Amazon to make Llama 2 available on Azure and AWS.
- Controversy surrounds Llama 2’s classification as open source due to licensing issues.
- Dell’s Validated Designs for Generative AI support a range of AI applications beyond inferencing.
- Deployment details for different Llama 2 models are provided by Dell.
Main AI News:
In a bold move, Dell has entered into a strategic partnership with Meta, the parent company of Facebook, to facilitate the seamless deployment of the Llama 2 large language model (LLM) on-premises, offering businesses an alternative to cloud-based access.
The enterprise landscape has seen a growing demand for companies to harness Meta’s AI prowess within their own IT infrastructure. Dell aims to position itself as the go-to provider for this essential kit, capitalizing on the rising need for on-premises AI capabilities.
At the heart of this collaboration lies Dell’s Validated Design for Generative AI portfolio, a collection of meticulously tested hardware configurations announced earlier this year, co-engineered with GPU giant Nvidia. In addition to these cutting-edge hardware solutions, Dell is extending its expertise by providing deployment and configuration guidance, ensuring rapid implementation and operationalization for its clients.
For instance, Dell has seamlessly integrated the Llama 2 models into its system sizing tools, offering customers invaluable insights into the optimal configuration tailored to their unique requirements.
Jeff Boudreau, Dell’s Chief AI Officer, emphasized the transformative potential of generative AI models like Llama 2, stating, “With the Dell and Meta technology collaboration, we’re making open source GenAI more accessible to all customers, through detailed implementation guidance paired with the optimal software and hardware infrastructure for deployments of all sizes.“
Llama 2, introduced in July, comes in various sizes, offering models with seven billion, 13 billion, and 70 billion parameters, each with distinct hardware demands. While it is freely available for research purposes, limited commercial use is also supported. Notably, Meta has already collaborated with tech giants Microsoft and Amazon to make Llama 2 accessible through the Azure and AWS cloud platforms.
However, there has been some debate about labeling Llama 2 as open source, given that it lacks approval from the Open Source Initiative (OSI) for its licensing terms.
Dell’s Validated Designs for Generative AI, unveiled in August, blends the company’s server infrastructure with Nvidia GPUs, storage solutions, and software like Nvidia’s AI Enterprise suite. These offerings come bundled with professional services to assist clients in launching generative AI solutions, albeit for a fee.
Initially tailored for inferencing tasks, spanning natural language generation, chatbots, virtual assistants, marketing, and content creation, Dell has since expanded the portfolio to support customization and fine-tuning of AI models.
According to Dell’s specifications, the Llama 2 model with seven billion parameters can be efficiently operated with just a single GPU. In contrast, the 13 billion parameter version necessitates two GPUs, and the 70 billion parameter variant demands eight GPUs. Detailed guidelines on deploying the 7 billion and 13 billion parameter models on the PowerEdge R760xa system are available, while the 70 billion parameter version requires a robust server like the PowerEdge XE9680 due to the eight-GPU requirement.
Conclusion:
The collaboration between Dell and Meta to enable on-premises deployment of Llama 2 AI signifies a strategic move in the market, offering businesses an alternative to cloud-based AI solutions. With Dell’s hardware expertise and Meta’s cutting-edge technology, this partnership addresses the rising demand for customizable AI models and deployment options. It enhances accessibility to generative AI, potentially transforming how industries operate and innovate, making it a noteworthy development in the AI market.
