Applying TorchTPU optimizations to FLUX.1-dev reduced image generation time from 11.49 seconds to 2.78 seconds per step.

From the articleSwitch Your PyTorch Model From Nvidia to Google TPUs With One Line: TorchTPU Is Now Open SourceAlso in Google Open-Sources TorchTPU, a Native PyTorch Backend for Cloud TPUs, Switch Your PyTorch Model From Nvidia to Google TPUs With One Line: TorchTPU Is Now Open Source, Google Open-Sources TorchTPU, a Native PyTorch Backend for Cloud TPUs