AWS Certified AI Practitioner Practice Exam

When training a large language model (LLM) with a concern for environmental impact, which Amazon EC2 instance type is preferred?

1. Amazon EC2 C series

2. Amazon EC2 G series

3. Amazon EC2 P series

4. Amazon EC2 Trn series

The correct answer is: Amazon EC2 Trn series

When considering the environmental impact of training a large language model, the Amazon EC2 Trn series is preferred because it is specifically designed for training machine learning models. These instances utilize AWS's custom-built Trainium chips, which are optimized for high-performance training of deep learning models. The Trainium processors provide superior performance per watt, which means they can complete training tasks more efficiently than other instance types, resulting in lower energy consumption. Choosing the Trn series not only enhances the speed of training large language models but also reduces the carbon footprint associated with the computational resources consumed. This aligns with strategies aimed at minimizing environmental impact while leveraging advanced machine learning capabilities. In contrast, other instance types such as the C series, G series, and P series may not offer the same level of efficiency for training deep learning models specifically, as they are designed for different workloads. The C series is optimized for compute-intensive workloads, the G series for graphics-intensive applications, and the P series for GPU-accelerated tasks, but they do not center their design around energy efficiency for model training like the Trn series does.