Pull latest changes for TensorFlow (#3298)
* Pull latest changes for TensorFlow
* Replace xla::int64_t with int64_t
* Replace xla::int16 with int16_t
* Replace xla::uint64 with uint64_t
* Replace xla::uint8 with uint8_t
* Replace xla::int32 with int32_t
* Replace xla::uint32 with uint32_t
* Replace xla::uint16 with uint16_t
* Replace xla::int8 with int8_t
* Update ArgMin/Max and fix import for AsInt64Slice
* Merge with latest changes
* Fix linter issues
* Remove multi_load_libtpu.diff
* Replace xla::int64_t with int64_t
* Replace xla::string with std::string
* Fix more linter issues
* Fix build issues in xla_client
* Fix linter issues
* Replace xla::int64_t with int64_t for latest commit
* Fix linter issues
* Pull latest TensorFlow again
