TL;DR:
- Deci introduces two groundbreaking AI Foundation Models: DeciDiffusion 1.0 and DeciLM 6B.
- Accompanying these models is the Infery LLM SDK, enhancing performance and cost-efficiency.
- Deci’s innovations reduce inference compute costs by up to 80%, enabling enterprise-grade integration.
- Models like DeciLM 6B offer exceptional speed while maintaining accuracy.
- Accessibility is a key feature, allowing self-hosting of models for customization and data privacy.
- The technology is set to transform the generative AI landscape across industries.
Main AI News:
Deci, the leading deep learning company dedicated to advancing AI’s capabilities, has introduced two revolutionary foundation models: DeciDiffusion 1.0 and DeciLM 6B, accompanied by the powerful Infery LLM SDK. This triumvirate of innovations is poised to establish new benchmarks for both performance and cost-efficiency within the domain of generative AI.
The ever-increasing computational demands inherent in training and deploying generative AI models have, until now, been a formidable obstacle to cost-effective integration and scalability. Deci’s pioneering solutions, however, bridge this gap, ushering in an era of efficient, cost-effective scaling that is tailor-made for seamless enterprise integration. With the utilization of Deci’s open-source generative models and the Infery LLM SDK, AI teams can anticipate a substantial reduction of up to 80% in inference compute costs, all while harnessing the capabilities of widely available and budget-friendly GPUs, including the NVIDIA A10. Furthermore, this transformative technology enhances the quality of AI applications across a spectrum of domains, encompassing content and code generation, image creation, chat applications, and beyond.
Among the array of models introduced by Deci, ‘DeciDiffusion 1.0’ emerges as a standout performer, delivering blazing-fast text-to-image generation that astoundingly produces high-quality images in less than a second, a feat achieved at three times the speed of the renowned Stable Diffusion 1.5 model. Stepping into the spotlight next is DeciLM 6B, boasting an impressive 5.7 billion parameters. While its accuracy rivals industry giants such as LLaMA 2 7B, Falcon-7B, and MPT-7B, what truly distinguishes it is its remarkable inference speed, clocking in at a mind-boggling 15 times faster than the Meta LLaMA 2 7B. Completing this formidable lineup is ‘DeciCoder,’ a 1 billion parameter code generation LLM, introduced just weeks ago. Notably, these models not only deliver unmatched inference speed but also maintain equivalent or superior accuracy.
Yonatan Geifman, CEO and co-founder of Deci, emphasizes the pivotal importance of mastering model quality, the inference process, and cost efficiency for generative AI’s transformative potential. He asserts, “At Deci, our journey and extensive collaborations with the world’s AI elite have equipped us to craft a solution that’s nothing short of transformative for enterprises diving into Generative AI. With our robust array of open-source models and cutting-edge tools, we’re setting the stage for teams to redefine excellence in their generative AI ventures.”
These models, meticulously developed through Deci’s proprietary Neural Architecture Search (AutoNAC™) technology, stand as exemplars of efficiency and effectiveness within today’s generative AI market. Accompanying the Foundation Models is the introduction of Infery LLM, an inference SDK that empowers developers to achieve a significant performance boost with existing LLMs while preserving the desired accuracy. The unprecedented efficiency unleashed by combining Deci’s open-source models and Infery LLM is underpinned by unique features such as continuous batching, advanced selective quantization, and ultra-efficient beam search, among others.
Deci’s offerings also break new ground in terms of accessibility. Unlike closed-source API models, Deci grants unrestricted access to models that can be self-hosted anywhere. This not only ensures superior performance and substantial reductions in inference costs at scale but also empowers users with customization options while alleviating concerns related to data privacy and compliance.
Prof. Ran El Yaniv, Chief Scientist and co-founder of Deci, underscores the critical importance of rapid, cost-effective AI implementation in today’s evolving landscape. He states, “With Deci’s groundbreaking solutions, companies receive both enterprise-grade quality and control, as well as the flexibility to customize models and the inference process according to their precise requirements. This commitment ensures unmatched excellence and a lasting competitive edge.”
Conclusion:
Deci’s latest generative AI innovations signify a transformative leap in performance and cost efficiency. These developments are poised to reshape the market by enabling businesses to harness the power of generative AI more effectively, resulting in rapid, cost-efficient AI implementation across diverse industries.
