[ROCm] Fix: Conditionally import torch.distributed to support builds without distributed

[0xDELUXA]

What does this PR do?

torch.distributed submodules are not available in all PyTorch builds (for example, Windows ROCm). Importing them unconditionally at the top level causes an immediate crash with:

ModuleNotFoundError: No module named 'torch._C._distributed_c10d'; 'torch._C' is not a package

This makes import torchao fail entirely on those builds, even for users who have no intention of using distributed training.

Prior art

The same fix was applied across the HuggingFace ecosystem:

• [core] conditionally import torch distributed stuff. huggingface/diffusers#12420
• fix more torch.distributed imports huggingface/diffusers#12425
• Fix error on importing unavailable torch.distributed huggingface/transformers#40038
• Fix: Conditionally import torch.distributed.fsdp in trainer_seq2seq.py huggingface/transformers#44507
• Fix: Conditionally import torch.distributed.algorithms.join in accelerator.py huggingface/accelerate#3962

Related issues

• [Issue]: PyTorch on Windows does not work with diffusers and torchao ROCm/ROCm#5906
• RDNA2 GPUs/iGPUs are not being enumerated with ROCm on Windows + Adrenalin 26.1.1 ROCm/ROCm#5871 (comment)
• [Feature]: torch.distributed support on Windows (workaround available for ComfyUI) ROCm/ROCm#5689
• Diffusers crash due to hard-dependency on torch._C._distributed_c10d huggingface/diffusers#12409

cc @danielvegamyhre @BowenBao

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/4017

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[meta-cla]

Hi @0xDELUXA!

Thank you for your pull request and welcome to our community.

Action Required

In order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you.

Process

In order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with CLA signed. The tagging process may take up to 1 hour after signing. Please give it that time before contacting us about it.

If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks!

[meta-cla]

Thank you for signing our Contributor License Agreement. We can now accept your code for this (and any) Meta Open Source project. Thanks!

[meta-cla]

Thank you for signing our Contributor License Agreement. We can now accept your code for this (and any) Meta Open Source project. Thanks!

[0xDELUXA]

Tracked down all issues preventing import torchao from working on Windows ROCm. This ended up being quite a large set of changes, and I don’t think it’s feasible to get this merged. Closing.