Use linux.4xlarge.memory instead of linux.12xlarge

[huydhn]

linux.4xlarge.memory has 128 GB of memory with 16 CPU cores while linux.12xlarge, a more expensive runner, has only 96 GB of memory on a whopping 48 CPU cores.

Testing

Example runs for:

• dl3 https://github.com/pytorch/executorch/actions/runs/11863238941/job/33064335597?pr=6896 (20m vs 19m in trunk)
• emformer_predict https://github.com/pytorch/executorch/actions/runs/11863238941/job/33064336524?pr=6896 (1h30m vs 1h20m in trunk)

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/6896

• 📄 Preview Python docs built from this PR

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

✅ You can merge normally! (1 Unrelated Failure)

As of commit 8204737 with merge base ec68eb3 ():

BROKEN TRUNK - The following job failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• trunk / test-coreml-delegate / macos-job (gh) (trunk failure)
  RuntimeError: Command bash /Users/runner/work/_temp/exec_script failed with exit code 1

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

[guangy10]

@huydhn Before merging can you link the comparison of job execution time between using linux.4xlarge.memory and linux.12xlarge? Would like to understand the actual trade-off we made, ideally we should expect it to be tiny

[guangy10]

Can you attach the job link to this model since we re-enable it ?

[huydhn]

Good catch, I think I will comment it out and leave it for latter. It doesn't OOM but take forever to export (close to 6 hours so far). I don't have much context, so probably need help from @dvorjackz to figure this one out