- Try Google Colaboratory’s TPU - Qiita
- How to use TPU from Keras with Google Colab and the speed difference with GPU - Diary of a Mogul, a Hippo and Powder
- Don’t just switch to TPU, convert the model
- Revert to CPU model when predict
- Learning transformer with TPU in Google Colab - Language and Intelligence
This page is auto-translated from /nishio/TPU using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.