Enable nightly build to test 2.14 - #3955
Merged
Merged
Conversation
Branch-cut step to start testing the 2.14 release in tutorials, mirroring #3929 for 2.13. The push-on-main trigger in build-tutorials-nightly.yml is already enabled, so only the pin moves this time.
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/tutorials/3955
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 689fd38 with merge base d445c1f ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
karthickai
approved these changes
Sep 2, 2026
atalman
added a commit
that referenced
this pull request
Sep 2, 2026
…3960) Fixes: #3816 `beginner_source/introyt/autogradyt_tutorial.py` fails on torch 2.14: ``` File "beginner_source/introyt/autogradyt_tutorial.py", line 493, in <module> with torch.autograd.profiler.profile(use_cuda=run_on_gpu) as prf: TypeError: profile.__init__() got an unexpected keyword argument 'use_cuda' ``` `use_cuda` was the deprecated alias for `use_device` and has been removed in 2.14: | ref | `profile.__init__` | | --- | --- | | `v2.13.0` | `use_cuda=False, # Deprecated` alongside `use_device=None` | | `v2.14.0-rc10` | `use_cuda` gone entirely; `use_device=None` remains | Switch to the supported spelling: ```diff -with torch.autograd.profiler.profile(use_cuda=run_on_gpu) as prf: +with torch.autograd.profiler.profile(use_device="cuda" if run_on_gpu else None) as prf: ``` `run_on_gpu` is already computed a few lines above from `torch.cuda.is_available()`, so behaviour is unchanged on both CPU and GPU runners. ## Where this is failing Surfaced by the nightly lane once #3955 moved it from 2.13.0 to 2.14.0 -- [failing run](https://github.com/pytorch/tutorials/actions/runs/33655943733/job/100334388632). `build-tutorials-nightly.yml` sets `USE_NIGHTLY=1`, so `.jenkins/build.sh` installs `torch==2.14.0` from the test channel. It will also start failing on the regular PR/push gate once #3959 lands, since that moves `.ci/docker/requirements.txt` from `torch==2.13` to `torch==2.14`. `.jenkins/get_files_to_run.py` shards every tutorial across the workers, so this file runs on every build. Standalone so it can land ahead of, and independently of, #3955 / #3959. ## Note on the trailing katex traceback The `sphinxcontrib.katex` `FileNotFoundError: '/tmp/tmp...'` at the end of that log is secondary -- Sphinx raises the gallery `ExtensionError` first, then the katex `build-finished` handler trips over an already-removed temp dir while unwinding. It should disappear with this fix. It is unrelated to the `yarn global add katex` pin in #3959. AI assistance (Claude) was used for this change.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Branch-cut step to start testing the 2.14 release in tutorials. Mirrors #3929 ("Enable nightly build to test 2.13").
.jenkins/build.sh: pintorch==2.14.0(from thetestchannel, cu130) on theUSE_NIGHTLYpath and the commented reference line; update the2.13 -> 2.14comment.Unlike #3929, no workflow change is needed: the
pushonmaintrigger in.github/workflows/build-tutorials-nightly.ymlis still enabled from the 2.13 cycle, so the nightly tutorials build already runs onmainand will pick up the 2.14 RC binaries with this pin..ci/docker/requirements.txtis intentionally left attorch==2.13— that gets bumped to 2.14 at the later "move to stable binaries" step, once 2.14 is released.Verification
torch==2.14.0+cu130is published on the test channel, so the pin resolves:cu130 remains the right channel for 2.14: CUDA 13.0 is the stable CUDA version in the 2.14 release matrix.
Release playbook reference (2.13 -> 2.14)