Contributed to the pytorch/pytorch repository by developing a unified device checking feature within the common_dtensor module, leveraging the Accelerator API to enhance cross-device compatibility. Focused on integrating accelerator-driven workflows, the work improved the efficiency of the testing module and established a foundation for broader accelerator support in future releases. The implementation utilized Python and distributed systems concepts, emphasizing robust testing practices to ensure device-agnostic functionality. By aligning common_dtensor with the Accelerator API, the contribution addressed the need for seamless operation across diverse hardware, supporting PyTorch’s ongoing efforts to expand compatibility and streamline development for distributed environments.