AMD And Intel Collaborate To Combine x86 CPU Optimization Work For PyTorch In oneDNN, Open-Source Deep Learning Library Governed By Unified Acceleration Foundation