Add a small cache for `xla::HloSharding` to `xla::ifrt::HloShardingSpec`
This will be useful once `xla::ifrt::HloShardingSpec` starts caching various sharding-related information. PiperOrigin-RevId: 972529987
J
Junwhan Ahn committed
ee6d552459b01e66ad8ad76304c15c130900444a
Parent: 0eaaeb4
Committed by jax authors <google-ml-automation@google.com>
on 8/28/2026, 12:33:11 PM