Dell is expanding its generative AI product portfolio to speed up the business transformation process. The company says its new on-premises tools are geared around intelligent data collection, multi-cloud access and professional services. This includes a validated design for Nvidia to improve model customisation and more data lakehouse solutions for multi-cloud AI.
The validated design technology for Nvidia devices includes a series of pre-trained models built to extract information from large datasets without a company having to build a model from scratch. It includes best practice techniques for fine-tuning generative AI models and is information secure as it runs on-premises.
These techniques and models work across a range of tasks and are effectively a scalable blueprint for customisation, Dell says. They can be used across proprietary data for training through diffusion, transfer learning and prompt tuning.
“To maximise AI efforts and support workloads across public clouds, on-premises environments and at the edge, companies need a robust data foundation with the right infrastructure, software and services,” said Jeff Boudreau, chief AI officer at Dell. “That’s what we are building with our expanded validated designs, professional services, modern data lakehouse and the world’s broadest GenAI solutions portfolio.”
As well as model tuning, the Dell Validated Designs for Generative AI has also been updated to include inferencing, where logical rules are applied to the knowledge base to evaluate and analyse new information. This, according to Dell, will allow users to more quickly deploy GenAI models with proven infrastructure – including its own AI server the Dell PowerEdge XE9680, which can be built with a choice of Nvidia chips.
Dell’s data storage and compute
Dell says its ability to combine on-premises compute power with storage options allows end users to rapidly feed and train models on secure company data across a range of storage and data types using validated designs.
The company is also using its own understanding of AI to improve customer performance and drive business results. Its professional services team are providing a range of services including data preparation, where customers are presented with clean, accurate data sets in the right format to get the most out of generative AI models.
Once the data has been modified for AI use Dell says it can also help a company establish an operational GenAI platform it can then use for inferencing and model customisation, which speeds up the time to value. The company also offers an educational services scheme to provide employees with critical AI skills.
A study into enterprise use of generative AI by TECHnalysis Research found that organisations made being able to use their own data to customisation foundation models a fundamental requirement of any implementation. “They also they need assistance in helping prep their data for that work,” said Bob O’Donnell, president and chief analyst, TECHnalysis Research. This included data cleaning and set-up.
Read more: How real is the threat of data poisoning to generative AI?
