Google Releases Open-Source TorchTPU to Run PyTorch Apps on TPUs
Summary
Google released TorchTPU, an open-source version of PyTorch that runs AI apps on its Tensor Processing Units with a configuration change and no rewriting. Amazon released similar software, TorchNeuron, for its Trainium chips a day earlier. Both companies want to loosen Nvidia's hold on AI development, where PyTorch is used by most AI apps.
Key Points
- Google says it worked with Meta, Hugging Face, Inferact and RadixArk to make TorchTPU work with their development tools.
- Google launches a TPU agent suite to help developers move apps to its chips and fix errors.
- Google updates its LiteRT tool so on-device AI apps built with PyTorch run better on phones and laptops.