Overview / Usage

The problem statement revolves around the need for efficient and accurate object detection in autonomous vehicles. Object detection is crucial for ensuring the safe navigation and operation of autonomous vehicles. It involves identifying and recognizing various objects such as vehicles, traffic signals, traffic signs, and pedestrians in real-time. Real-time processing with high accuracy is essential for timely decision-making and avoiding potential hazards on the road. Improving the efficiency and speed of object detection algorithms is crucial to enhance the overall performance and reliability of autonomous vehicles. Addressing these challenges and optimizing object detection algorithms is the key objective of this project

Methodology / Approach

Our solution focuses on enhancing the performance of a deep learning model by leveraging Intel oneAPI libraries, including the Intel Extension for PyTorch, Intel Neural Compressor, and Intel oneVPL. By integrating these libraries, we optimize the model's training time, reduce inference latency, and achieve low latency. The Intel Extension for PyTorch enables seamless integration with Intel hardware accelerators, while the Neural Compressor compresses and prunes the model for efficiency. Additionally, the utilization of Intel oneVPL accelerates video and image processing tasks during inference. By harnessing the power of Intel's advanced technologies, our solution maximizes model performance, leading to improved efficiency and superior outcomes.

Technologies Used

streamlit - webapp

**one api-**for optimize and compress

**devcloud-**train a model

​

Repository

https://github.com/vishnu2909200/autonomous_-vehicles-_oneapi.git