Abstract: In recent years, transformer models based on attention mechanisms have been widely adopted in natural language processing and computer vision. However, their high computational complexity ...

Overview Present-day serverless systems can scale from zero to hundreds of GPUs within seconds to handle unexpected increases ...

Abstract: In this paper, a hybrid model combining Convolutional Neural Network (CNN) and Transformer architectures is proposed for wind speed forecasting. In this model, the CNN is employed to extract ...