80
/
2
Min Read
Google has released TorchTPU, its version of PyTorch, as an open-source project, joining Amazon in trying to loosen Nvidia’s hold on AI development. PyTorch is the widely used open-source toolkit for building, training and running AI models, used by about 85% of AI apps by some accounts.
Meta developed it in 2016 and handed it to the Linux Foundation in 2022. Many of its features were built to run on Nvidia GPUs.
Google said it worked with Meta, Hugging Face, Inferact (maker of vLLM) and RadixArk (maker of SGLang) so TorchTPU works well with popular AI development tools. It also launched a TPU agent suite to help developers migrate and fix errors, and updated LiteRT, its tool for running AI on phones, laptops and embedded devices, to work better with PyTorch apps.
Amazon made a similar move on Monday with TorchNeuron, which ports AI apps to its Trainium chip and AWS without rewriting. Amazon is also submitting code to the main PyTorch project to decouple features from Nvidia’s chips.
The tools do not mean developers will leave Nvidia hardware en masse, but both companies are working to end Nvidia’s advantage.
This article was originally
published by the 
