News
Accelerate PyTorch with IPEX and oneDNN Using Intel BF16 Technology
February 3, 2021
Intel and Facebook previously collaborated to enable BF16, a first-class data type in PyTorch. It supports basic math and tensor operations and adds CPU optimization with multi-threading, vectorization, and neural network kernels from oneAPI Deep Neural Network Library (oneDNN, formerly known as MKL-DNN).