Analysis of CNN Model with Traditional Approach and Cloud AI based Approach
Utkarsh Kushwaha, Puja Gupta, Sonu Airen, Megha Kuliha · 2022 International Conference on Automation, Computing and Renewable Systems (ICACRS) · 2022
A plethora of work has been done in the field of AI, including research and real-time applications; however, cloud computing and its technologies have amended Artificial Intelligence applications; cloud computing has aided a lot in deep learning, which is a subset of AI; with the use of cloud computing, the performance and improvement of data set and model, the time and resource consuming process has become much more facile. Making data sets suitable for model training is a time-consuming operation; hence, a transfer learning idea is proposed for simplification, but it must be concrete to the desired data set. The advancement of cloud AI and the usage of resources such as the GPU made it possible to train and evaluate CNN models in less time. This study compares the performance and feasibility of CPU-based local computers and GPU-based cloud platforms like Google colaboratory in terms of speed, power, and security.