Comparative Analysis of Open-Source Deep Learning Frameworks for GPU-Accelerated Performance
Computer Science Engineering and Technology · 2024
Introduction: Artificial intelligence has been transformed by deep learning, which makes it possible to recognize patterns and do intricate calculations.Deep network training is computationally demanding, though.GPU acceleration is used for performance in a number of open-source deep learning frameworks, such as Tensor Flow, Py Torch, and CNTK.The capabilities, effectiveness, and applicability of these frameworks for various machine learning applications are compared and assessed in this study.Research signification: For researchers and developers, it is essential to comprehend the advantages and disadvantages of different deep learning frameworks.The best hardware-software combination for AI applications is chosen with the aid of this study.Additionally, it emphasizes how deep learning contributes to technological improvements in domains including image processing, medical diagnostics, and e-learning.Research methodology: Alternatives: Microsoft Cognitive Toolkit, Neuroph, Torch, Tensor Flow, Bit name Py torch.Evaluation parameters: Data Availability and Quality, Computational resources, Explain ability, Cost of Installation.Result: The results Torch achieved the highest rank, while Microsoft Cognitive Toolkit the lowest rank is attained.Torch has the highest value for Deep Learning according to the GRA approach.