The forthcoming wave of digital and data transformation is poised to revolve around generative AI (GenAI). Businesses are now emphasizing the development of AI, often requiring educational versions based on proprietary data. This shift raises significant inquiries regarding cost, security, and data ownership.
In response, there is a growing trend towards on-premises operations utilizing open-source large language models like Llama 2. These models ensure regulatory compliance by offering predictable expenses, data control, and reduced risks of security breaches and data leaks. To provide enhanced assistance to businesses, Dell and Meta have joined forces.
Transitioning from AI proof-of-concept to actual implementation can pose challenges for some companies. This is where the collaboration between Touching Experience and Dell comes into play.
By introducing a system that simplifies the selection, construction, and optimization of AI models tailored to specific business use cases using Dell’s infrastructure, Dell and Hugging Face are working together to streamline this process.
Collaboration with Hugging Face
Dell’s partnership with Hugging Face aims to streamline the development, optimization, and deployment of open-source conceptual AI models for businesses. Leveraging the Touching Experience community, this collaboration seeks to integrate Dell’s premier hardware products and services.
Central to this partnership is a new Dell platform built on the Hugging Face framework. Through this portal, Dell customers can swiftly choose from a range of open-source AI models optimized for efficiency and specific use cases.
The platform offers customized containers and scripts for deploying these models on Dell hardware and data storage systems. Additionally, it provides access to templates and models for specific outcomes, datasets, libraries, and tutorials for training conceptual AI models. Companies can combine these resources with their proprietary data to tailor the models to their needs.
Fine-tuning AI models can be a complex and time-consuming process, often requiring techniques like retrieval augmented generation (RAG), LoRA, and QLoRA. Dell employs containerized tools to enhance this process.
Customers will have immediate access to a variety of Dell PowerEdge servers designed for AI models. Dell’s APEX support and workstation assistance will be available soon.
This collaboration, leveraging the strengths of Dell’s infrastructure and Hugging Face’s AI capabilities, represents a significant advancement in enhancing the accessibility and security of open-source AI for enterprise applications.
Dell’s GenAI Validated Design with Llama Meta
Furthermore, Dell has introduced pre-tested hardware and software solutions for GenAI-related tasks known as the Dell Validated Design for GenerativeAI. Recently, Dell partnered with Meta to extend this validated design, making it easier to deploy Meta’s Llama 2 AI models on-premises.
The Dell Validated Design for Generative AI streamlines the implementation and management of on-premises GenAI projects with the pre-tested system of Llama 2. This Dell architecture, validated on Llama 2, offers configuration settings and deployment guidelines, simplifying the setup and operation of Llama 2 on Dell platforms.
The Dell PowerEdge XE9680 server, equipped with eight NVIDIA H100 GPUs and NVIDIA AI software, is ideal for deploying Llama 2. Additionally, Llama 2 can be integrated with various Dell systems, such as the PowerEdge R760xa.
Dell has explored the flexibility of GenAI models, demonstrating their application with Llama 2 versions. This includes running the Llama 2 70B model on a PowerEdge XE9680 with Ubuntu 22.0 and NVIDIA GPUs, yielding precise and valuable AI results.
To aid customers in selecting optimal AI solutions, Dell incorporates Llama 2 models into its internal tools. This comprehensive approach provides businesses with reliable GenAI options across desktops, data centers, and clouds, ensuring efficient deployment and management of on-premises GenAI projects.
Insights from an Analyst
Establishing a partnership with Meta is pivotal. Sy Choudhury, Meta’s vice president of AI partnerships, participated in a recent Dell AI-focused event in Austin. When questioned about the significance of this partnership for Meta, he emphasized the importance of maintaining an open path to conceptual AI.
This collaboration aligns with Meta’s partner-friendly approach, ensuring an ecosystem conducive to large language models. The collaboration between Dell and Hugging Face follows a similar pattern.
While each of Dell’s competitors pursues a distinct strategy, Dell stands out by offering tuned and validated technology solutions, base models, and streamlined tools, thereby alleviating the complexities associated with implementing GenAI for businesses of all sizes.
As conceptual AI continues to reshape the landscape of IT, Dell remains committed to facilitating the seamless adoption of AI technologies, as evidenced by its partnerships with Meta and Hugging Face, and its organizational adjustments to support enterprise AI. Dell’s initiatives in the realm of GenAI promise to benefit customers, simplifying the tasks of IT professionals and enhancing operational efficiency.
