-
|
Is JAX support for using the CPU built into the TPU board something that might happen in the future? Just out of curiosity, how was data loading handled for the large scale MLperf benchmarks? Is it currently like Pytorch-xla where you need a VM for each TPU board? |
Beta Was this translation helpful? Give feedback.
Answered by
bilal2vec
Dec 9, 2020
Replies: 1 comment 1 reply
-
|
given the awesome news from neurips today I think this has pretty much been answered :) |
Beta Was this translation helpful? Give feedback.
1 reply
-
|
Hopefully you'll forgive our silence up until now! We didn't want to spoil the surprise. |
Beta Was this translation helpful? Give feedback.
Answer selected by
bilal2vec
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
given the awesome news from neurips today I think this has pretty much been answered :)