[torch-xla 2.7] Change xm.xrt_world_size() to xr.world_size(). xm.get_ordinal() to xr.global_ordinal()

[jeffhataws]

System Info

In torch-xla version 2.7, xm.xrt_world_size() is replaced by xr.world_size() and xm.get_ordinal() is replaced by xr.global_ordinal() (xr is "import torch_xla.runtime as xr")

https://github.com/search?q=repo%3Ahuggingface%2Ftransformers+xrt_world_size&type=code
https://github.com/search?q=repo%3Ahuggingface%2Ftransformers+get_ordinal&type=code

Currently these show up as warnings in torch-xla 2.5/2.6:

WARNING:root:torch_xla.core.xla_model.xrt_world_size() will be removed in release 2.7. is deprecated. Use torch_xla.runtime.world_size instead.            
WARNING:root:torch_xla.core.xla_model.xla_model.get_ordinal() will be removed in release 2.7. is deprecated. Use torch_xla.runtime.global_ordinal instead. 

Who can help?

No response

Information

• The official example scripts
• My own modified scripts

Tasks

• An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• My own task or dataset (give details below)

Reproduction

1. Run any finetuning/training test on Neuron or TPU
2. Look for warnings like below: WARNING:root:torch_xla.core.xla_model.xrt_world_size() will be removed in release 2.7. is deprecated. Use torch_xla.runtime.world_size instead.
   WARNING:root:torch_xla.core.xla_model.xla_model.get_ordinal() will be removed in release 2.7. is deprecated. Use torch_xla.runtime.global_ordinal instead.

Expected behavior

No warnings.

With torch-xla 2.7, you will see errors.

[Rocketknight1]

This is in trainer so cc @muellerzr @SunMarc - I think we can just search-replace with the new methods?

[SunMarc]

Yeah, we are replacing them in accelerate. I'll ping the author to do the same in transformers