🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/155649

• 📄 Preview Python docs built from this PR
• 📄 Preview C++ docs built from this PR
• ❓ Need help or want to give feedback on the CI? Visit the bot commands wiki or our office hours

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

✅ No Failures

As of commit 89f38af with merge base 2259dbe ():
💚 Looks good so far! There are no failures yet. 💚

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

@svekars any update on this? Thanks!

Thank you both!

@svekars any blockers on this PR getting merged? Thanks! 🚀

@svekars @mikaylagawarecki are there any blockers I can take care of to get this PR merged?

@pytorchbot merge

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Merge failed

Reason: 9 mandatory check(s) failed. The first few are:

• pull / linux-jammy-cuda12.8-py3.10-gcc11-sm89 / build
• pull / linux-jammy-cuda12.8-py3.10-gcc11-test / test (distributed, 1, 3, lf.linux.g4dn.12xlarge.nvidia.gpu)
• pull / linux-jammy-cuda12.8-py3.10-gcc11-test / test (distributed, 2, 3, lf.linux.g4dn.12xlarge.nvidia.gpu)
• pull / linux-jammy-cuda12.8-py3.10-gcc11 / test (default, 1, 5, lf.linux.4xlarge.nvidia.gpu)
• pull / linux-jammy-cuda12.8-py3.10-gcc11 / test (default, 4, 5, lf.linux.4xlarge.nvidia.gpu)

Dig deeper by viewing the failures on hud

@svekars @mikaylagawarecki trying to sync with latest main to remove merge failures

Pull workflow has not been scheduled for the PR yet. It could be because author doesn't have permissions to run those or skip-checks keywords were added to PR/commits, aborting merge. Please get/give approval for the workflows and/or remove skip ci decorators before next merge attempt. If you think this is a mistake, please contact PyTorch Dev Infra.

@pytorchbot merge

@pytorchbot merge

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Merge failed

Reason: 1 jobs have failed, first few of them are: trunk / macos-py3-arm64 / build

@pytorchbot merge

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Merge failed

Reason: 1 jobs have failed, first few of them are: trunk / macos-py3-arm64 / test (default, 1, 3, macos-m1-stable)

@mikaylagawarecki @svekars is it possible to rerun the one failing test to get this green? I'm not sure how to do that. I don't see any documentation on a pytorchbot command for that.

Thank you for your help. Sorry a simple doc string fix is taking this much time and effort!

@loganthomas I keep getting an error that disappears when I run it. Will run it again and send it your way to see if you can make heads or tails of it. It has something to do with the example in the docstring

@loganthomas see error below:

=== Found 1 errors ===
--- Error: 1 / 1 ---
    * REASON: GotWantException
    DOCTEST DEBUG INFO
      XDoc "/Users/ec2-user/runner/_work/_temp/venv-3.12-1755177273/lib/python3.12/site-packages/torch/nn/modules/loss.py::CrossEntropyLoss:1", line 17 <- wrt doctest
      File "/Users/ec2-user/runner/_work/_temp/venv-3.12-1755177273/lib/python3.12/site-packages/torch/nn/modules/loss.py", line 1349, <- wrt source file
    DOCTEST PART BREAKDOWN
    Passed Parts:
         2 >>> # Example of target with incorrectly specified class probabilities
         3 >>> loss = nn.CrossEntropyLoss()
         4 >>> torch.manual_seed(283)
         5 >>> input = torch.randn(3, 5, requires_grad=True)
         6 >>> target = torch.randn(3, 5)
         7 >>> # Provided target class probabilities are not in range [0,1]
         8 >>> target
        13 >>> target.sum(axis=1)
    Failed Part:
        16 >>> loss(input, target).item()
    Remaining Parts:
        18 >>>
        19 >>> # Example of target with correctly specified class probabilities
        20 >>> # Use .softmax() to ensure true probability distribution
        21 >>> target_new = target.softmax(dim=1)
        22 >>> # New target class probabilities all in range [0,1]
        23 >>> target_new
        27 >>> # New target class probabilities sum to 1
        28 >>> target_new.sum(axis=1)
        30 >>> loss(input, target_new).item()
    DOCTEST TRACEBACK
    Expected:
        4.6379876136779785
    Got:
        4.637986660003662
    Repr Difference:
        got  = '4.637986660003662'
        want = '4.6379876136779785'
    DOCTEST REPRODUCTION
    CommandLine:
        python -m xdoctest /Users/ec2-user/runner/_work/_temp/venv-3.12-1755177273/lib/python3.12/site-packages/torch/nn/modules/loss.py CrossEntropyLoss:1

=== Failed tests ===
python -m xdoctest /Users/ec2-user/runner/_work/_temp/venv-3.12-1755177273/lib/python3.12/site-packages/torch/nn/modules/loss.py CrossEntropyLoss:1
=== 1 failed, 338 passed, 392 skipped, 153 warnings in 10.01 seconds ===
GITHUB_RUN_ID, GITHUB_RUN_ATTEMPT, or ARTIFACTS_FILE_SUFFIX not set, not uploading
Uploading artifacts took 0.00 seconds
Writing 1 documents to S3 ossci-raw-job-status/ossci_uploaded_metrics/td_test_failure_stats_v2_1755177625_70944228791111f08d22268327012cbd
Done! Finish writing document to S3 ossci-raw-job-status/ossci_uploaded_metrics/td_test_failure_stats_v2_1755177625_70944228791111f08d22268327012cbd 
    main()
  File "/Users/ec2-user/runner/_work/pytorch/pytorch/test/run_test.py", line 2037, in main
    run_tests(
  File "/Users/ec2-user/runner/_work/pytorch/pytorch/test/run_test.py", line 1872, in run_tests
    raise RuntimeError(failure.message + keep_going_message)
RuntimeError: doctests 1/1 failed!

Tip: You can keep running tests even on failure by passing --keep-going to run_test.py.
If running on CI, add the 'keep-going' label to your PR and rerun your jobs.
Error: Process completed with exit code 1.

@loganthomas see error below:

=== Found 1 errors ===
--- Error: 1 / 1 ---
    * REASON: GotWantException
    DOCTEST DEBUG INFO
      XDoc "/Users/ec2-user/runner/_work/_temp/venv-3.12-1755177273/lib/python3.12/site-packages/torch/nn/modules/loss.py::CrossEntropyLoss:1", line 17 <- wrt doctest
      File "/Users/ec2-user/runner/_work/_temp/venv-3.12-1755177273/lib/python3.12/site-packages/torch/nn/modules/loss.py", line 1349, <- wrt source file
    DOCTEST PART BREAKDOWN
    Passed Parts:
         2 >>> # Example of target with incorrectly specified class probabilities
         3 >>> loss = nn.CrossEntropyLoss()
         4 >>> torch.manual_seed(283)
         5 >>> input = torch.randn(3, 5, requires_grad=True)
         6 >>> target = torch.randn(3, 5)
         7 >>> # Provided target class probabilities are not in range [0,1]
         8 >>> target
        13 >>> target.sum(axis=1)
    Failed Part:
        16 >>> loss(input, target).item()
    Remaining Parts:
        18 >>>
        19 >>> # Example of target with correctly specified class probabilities
        20 >>> # Use .softmax() to ensure true probability distribution
        21 >>> target_new = target.softmax(dim=1)
        22 >>> # New target class probabilities all in range [0,1]
        23 >>> target_new
        27 >>> # New target class probabilities sum to 1
        28 >>> target_new.sum(axis=1)
        30 >>> loss(input, target_new).item()
    DOCTEST TRACEBACK
    Expected:
        4.6379876136779785
    Got:
        4.637986660003662
    Repr Difference:
        got  = '4.637986660003662'
        want = '4.6379876136779785'
    DOCTEST REPRODUCTION
    CommandLine:
        python -m xdoctest /Users/ec2-user/runner/_work/_temp/venv-3.12-1755177273/lib/python3.12/site-packages/torch/nn/modules/loss.py CrossEntropyLoss:1

=== Failed tests ===
python -m xdoctest /Users/ec2-user/runner/_work/_temp/venv-3.12-1755177273/lib/python3.12/site-packages/torch/nn/modules/loss.py CrossEntropyLoss:1
=== 1 failed, 338 passed, 392 skipped, 153 warnings in 10.01 seconds ===
GITHUB_RUN_ID, GITHUB_RUN_ATTEMPT, or ARTIFACTS_FILE_SUFFIX not set, not uploading
Uploading artifacts took 0.00 seconds
Writing 1 documents to S3 ossci-raw-job-status/ossci_uploaded_metrics/td_test_failure_stats_v2_1755177625_70944228791111f08d22268327012cbd
Done! Finish writing document to S3 ossci-raw-job-status/ossci_uploaded_metrics/td_test_failure_stats_v2_1755177625_70944228791111f08d22268327012cbd 
    main()
  File "/Users/ec2-user/runner/_work/pytorch/pytorch/test/run_test.py", line 2037, in main
    run_tests(
  File "/Users/ec2-user/runner/_work/pytorch/pytorch/test/run_test.py", line 1872, in run_tests
    raise RuntimeError(failure.message + keep_going_message)
RuntimeError: doctests 1/1 failed!

Tip: You can keep running tests even on failure by passing --keep-going to run_test.py.
If running on CI, add the 'keep-going' label to your PR and rerun your jobs.
Error: Process completed with exit code 1.

@mikaylagawarecki this looks like a precision issue maybe something that isnt such a big issue?

@sekyondaMeta thank you! Yes, this appears to be a minor precision discrepancy and not blocking in the context of the example. For the main purpose of the example, these should be considered equivalent.

Yea, I think you can either update the value or skip the doctest for that example by doing something like

@sekyondaMeta @mikaylagawarecki thank you both. is there a way to kick off a new test run? happy to do this is there is a build int command.

we have to kick it off for you, but I just did that :)

@pytorchbot merge

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

@sekyondaMeta @mikaylagawarecki @svekars 🎉 thank you so much for all your help!!!