Samsung showed off the next iteration of its multimodal generative AI model, Gauss2, at the Samsung Developer Conference 2024. The model will be available in three sizes - Compact, Balanced, and Supreme.
Compact is a small-sized model built for speed and efficiency and is ideal for on-device usage "by maximizing the utilization of the device's computing resources."
Balanced offers a mix of performance and efficiency aimed at diverse tasks that require consistency. Supreme is built for performance, and Samsung says it reduces "computational costs during training and inference processes while maintaining high levels of both performance and efficiency."
The model supports up to 14 languages and various programming languages as well. Samsung uses a "custom tokenizer" for efficiency and performance in supported languages. The company says the model's processing speed "per hour is 1.5 to 3 times faster," compared to open-source generative AI models.
Samsung said that its in-house coding assistant, 'code.i,' is now being utilized by 60% of all software developers within the company.
Samsung says it will apply the Gauss2 model to increase productivity within the company. It "will continue to expand the reach of its AI-based services across all product lines so that users can experience more convenient and enjoyable daily lives."
Are you a techie who knows how to write? Then join our Team! Wanted:
- News Writer (Romania based)
Details here