Amazon's TorchNeuron lets PyTorch code run on Trainium chips with one line
Amazon Web Services released TorchNeuron, an open source native PyTorch backend that lets existing PyTorch code run on its Trainium AI chips with a one line device change. It is the most direct attack yet on the software lock in that keeps most AI developers on Nvidia GPUs, but the first release is a public beta with real gaps.