Is XLA feature existed in current repo?

[geekan]

❓ Questions & Help

I find https://news.developer.nvidia.com/nvidia-achieves-4x-speedup-on-bert-neural-network/ says tensorflow XLA has higher speed on bert, however, the pull request in this repo it mentioned #116 didn't implement something like XLA. Is the XLA feature already exist?

[LysandreJik]

Hi, there is no XLA/TPU support with the current library. Maybe in a future release!