Het Shah

@GleasonK Thanks so much for looking into this! Regarding (2), could you give me a code pointer to where PTXLA is currently building the V1 annotations? Also, regarding (1), we...

Hey everyone, sorry for the long message. I'm working on adding OpShardingV2 annotations to torch/xla (mentioned by @GleasonK as work item 2 [here](https://github.com/pytorch/xla/issues/9348#issuecomment-2970907670)). I found the following documentation that gives...

Thanks for the responses everyone! Sorry for the confusion regarding example 4. I had edited the comment shortly after I posted it once I realized the typo, but it seems...

Hey everyone, I added a PR for shardy support: https://github.com/pytorch/xla/pull/9541. Could folks please look at it for correctness when you get a chance? Want to make sure I haven't missed...

Hey all, could I get some eyes on the above PR please: https://github.com/pytorch/xla/pull/9541?

Very excited for this! Was wondering what the plan would be to support SPMD flows? Currently we use Torch-XLA's `xs.mark_sharding()` APIs to shard input tensors, and later use Shardy propagation...

Gotcha, thx for the response @cjonesy20. So I guess SPMD isn't going to be supported in the first release of TorchTPU (since IIRC torch.distributed doesn't explicitly support SPMD)? Also, is...

Any updates on this issue? We're trying to use torch-xla SPMD mode to run inference on gpt-oss and any help (whether that's example code, or just documentation) would be greatly...