PyTorch 2.13 brings FlexAttention to Apple Silicon, cuts peak memory by up to 4× for large-vocabulary models with nn.LinearCrossEntropyLoss, and updates distributed training, compilation, profiling, and on-device inference.Our live Q&A examined CUDA version support, CuTeDSL in TorchInductor, Python 3.15, ExecuTorch, torch.compile, and plans for PyTorch 2.14.Andrey Talman (@Meta), @albanDesmaison (@Meta), and Piotr Bialecki (@NVIDIA) joined moderator Chris Gottbrath (Gottbrath Technologies).This

Coding up a basic multi-layer perceptron in PyTorch. Introducing nn.Module feed-word pattern. Why is that cool? we can use the same 2- 2- ReLU- 1 structure to build an XOR logic gate, which can't be built without an MLP. https://t.co/97jfTszXo8