Refactor autocast C++ APIs to be device-agnostic #124359

guangyey · 2024-04-18T05:13:46Z

Stack from ghstack (oldest at bottom):

Motivation

This PR aims to refactor autocast C++ APIs to be device-agnostic and deprecate the device-specific autocast C++ APIs.
In C++ side,

is_enabled() -> is_enabled(device_type).
set_enabled(new_enabled) -> set_enabled(device_type, new_enabled).
get_autocast_dtype() -> get_autocast_dtype(device_type)
set_autocast_dtype(dtype) -> set_autocast_dtype(device_type, dtype)

These following C++ APIs are deprecated and should be removed in PyTorch 2.5

is_cpu_enabled
set_cpu_enabled
get_autocast_cpu_dtype
set_autocast_cpu_dtype
is_xpu_enabled
set_xpu_enabled
get_autocast_xpu_dtype
set_autocast_xpu_dtype
is_ipu_enabled
set_ipu_enabled
get_autocast_ipu_dtype
set_autocast_ipu_dtype
is_hpu_enabled
set_hpu_enabled
get_autocast_hpu_dtype
set_autocast_hpu_dtype
is_xla_enabled
set_xla_enabled
get_autocast_xla_dtype
set_autocast_xla_dtype
is_privateuseone_enabled
set_privateuseone_enabled
get_autocast_privateuseone_dtype
set_autocast_privateuseone_dtype

In Python side,
provide 4 generic autocast APIs:

torch.is_autocast_enabled(device_type)
torch.set_autocast_enabled(device_type, new_enabled)
torch.get_autocast_dtype(device_type)
torch.set_autocast_dtype(device_type, dtype)

Additional Context

We will submit another PR to refactor autocast Python APIs based on this PR.

cc @mcarilli @ptrblck @leslie-fang-intel @jgong5 @voznesenskym @penguinwu @EikanWang @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @chauhang

pytorch-bot · 2024-04-18T05:13:49Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/124359

📄 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.

✅ You can merge normally! (1 Unrelated Failure)

As of commit 927c0b4 with merge base 7e095be ():

FLAKY - The following job failed but was likely due to flakiness present on trunk:

periodic / win-vs2019-cuda11.8-py3 / test (default, 1, 4, windows.g5.4xlarge.nvidia.gpu) (gh)
test_linalg.py::TestLinalgCUDA::test_matmul_small_brute_force_1d_Nd_cuda_float32

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

ghstack-source-id: eeb80ac Pull Request resolved: #124359

guangyey · 2024-04-18T10:37:54Z

@albanD This PR is working in progress. Is this PR reasonable as it is BC-breaking? If it is reasonable, I will continue finishing it. Could you give me your input? Thanks.

[ghstack-poisoned]

cc mcarilli ptrblck leslie-fang-intel jgong5 [ghstack-poisoned]

albanD · 2024-04-18T17:01:28Z

I definitely love the direction!
We do want to be a bit careful with public API changes though:

In c++ if a github search doesn't return any hit, then I'm happy with making the BC-breaking change in this PR.
In python, we should deprecate all APIs for at least one version (if mostly un-used) or two version (if we see significant use via github search). The full process would be:
- This PR can add the new API
- This PR can add a warning about deprecation for the old API
- For APIs only needing one version, just after the 2.4 branch cut, we can do a follow up deleting the deprecated APIs
- For APIs needing two versions, wait after the 2.5 branch cut for the follow up.

cc mcarilli ptrblck leslie-fang-intel jgong5 [ghstack-poisoned]

# Motivation This PR aims to refactor autocast APIs to be device-agnostic and remove the device-specific autocast APIs. In C++ side, `is_enabled()` -> `is_enabled(device_type)`. `set_enabled(new_enabled)` -> `set_enabled(device_type, new_enabled)`. `get_autocast_dtype()` -> `get_autocast_dtype(device_type)` `set_autocast_dtype(dtype)` -> `set_autocast_dtype(device_type, dtype)` ~~`is_cpu_enabled`~~ ~~`set_cpu_enabled`~~ ~~`get_autocast_cpu_dtype`~~ ~~`set_autocast_cpu_dtype`~~ ~~`is_xpu_enabled`~~ ~~`set_xpu_enabled`~~ ~~`get_autocast_xpu_dtype`~~ ~~`set_autocast_xpu_dtype`~~ ~~`is_ipu_enabled`~~ ~~`set_ipu_enabled`~~ ~~`get_autocast_ipu_dtype`~~ ~~`set_autocast_ipu_dtype`~~ ~~`is_hpu_enabled`~~ ~~`set_hpu_enabled`~~ ~~`get_autocast_hpu_dtype`~~ ~~`set_autocast_hpu_dtype`~~ ~~`is_xla_enabled`~~ ~~`set_xla_enabled`~~ ~~`get_autocast_xla_dtype`~~ ~~`set_autocast_xla_dtype`~~ ~~`is_privateuseone_enabled`~~ ~~`set_privateuseone_enabled`~~ ~~`get_autocast_privateuseone_dtype`~~ ~~`set_autocast_privateuseone_dtype`~~ In Python side, provide 4 generic autocast APIs: -`torch.is_autocast_enabled(device_type)` -`torch.set_autocast_enabled(device_type, new_enabled)` -`torch.get_autocast_dtype(device_type)` -`torch.set_autocast_dtype(device_type, dtype)` cc mcarilli ptrblck leslie-fang-intel jgong5 [ghstack-poisoned]

guangyey · 2024-04-19T00:58:17Z

I definitely love the direction! We do want to be a bit careful with public API changes though:

In c++ if a github search doesn't return any hit, then I'm happy with making the BC-breaking change in this PR.

In python, we should deprecate all APIs for at least one version (if mostly un-used) or two version (if we see significant use via github search). The full process would be:

This PR can add the new API

This PR can add a warning about deprecation for the old API

For APIs only needing one version, just after the 2.4 branch cut, we can do a follow up deleting the deprecated APIs

For APIs needing two versions, wait after the 2.5 branch cut for the follow up.

Thanks for your elaboration. I capture your point.

guangyey · 2024-04-23T01:48:22Z

torch/_functorch/_aot_autograd/utils.py

-        torch.is_autocast_cpu_enabled(),
-        torch.get_autocast_gpu_dtype(),
-        torch.get_autocast_cpu_dtype(),
+        torch.is_autocast_enabled("cuda"),


fix UT warnings check issue on xla/test/test_dynamo.py::test_all_cpu_tensor

guangyey · 2024-04-23T08:00:56Z

@pytorchbot merge

pytorchmergebot · 2024-04-23T08:02:42Z

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

Advanced Debugging

Check the merge workflow status
here

# Motivation This PR aims to refactor autocast **C++** APIs to be device-agnostic and deprecate the device-specific autocast **C++** APIs. In C++ side, - `is_enabled()` -> `is_enabled(device_type)`. - `set_enabled(new_enabled)` -> `set_enabled(device_type, new_enabled)`. - `get_autocast_dtype()` -> `get_autocast_dtype(device_type)` - `set_autocast_dtype(dtype)` -> `set_autocast_dtype(device_type, dtype)` These following C++ APIs are deprecated and should be removed in PyTorch 2.5 - `is_cpu_enabled` - `set_cpu_enabled` - `get_autocast_cpu_dtype` - `set_autocast_cpu_dtype` - `is_xpu_enabled` - `set_xpu_enabled` - `get_autocast_xpu_dtype` - `set_autocast_xpu_dtype` - `is_ipu_enabled` - `set_ipu_enabled` - `get_autocast_ipu_dtype` - `set_autocast_ipu_dtype` - `is_hpu_enabled` - `set_hpu_enabled` - `get_autocast_hpu_dtype` - `set_autocast_hpu_dtype` - `is_xla_enabled` - `set_xla_enabled` - `get_autocast_xla_dtype` - `set_autocast_xla_dtype` - `is_privateuseone_enabled` - `set_privateuseone_enabled` - `get_autocast_privateuseone_dtype` - `set_autocast_privateuseone_dtype` In Python side, provide 4 generic autocast APIs: - `torch.is_autocast_enabled(device_type)` - `torch.set_autocast_enabled(device_type, new_enabled)` - `torch.get_autocast_dtype(device_type)` - `torch.set_autocast_dtype(device_type, dtype)` # Additional Context We will submit another PR to refactor autocast **Python** APIs based on this PR. cc mcarilli ptrblck leslie-fang-intel jgong5 voznesenskym penguinwu EikanWang Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang [ghstack-poisoned]

# Motivation Refactor autocast usage scenario in `torch/amp/autocast_mode.py` and `torch/utils/checkpoint.py` to fix the bug - convention conflict between `torch.xxx.get_autocast_xxx_dtype` defined in `autocast_mode.py` and `torch.xxx.get_autocast_dtype` defined in `checkpoint.py`. # Solution Use device-agnostic APIs like `torch.get_autocast_dtype`, ..., instead. Pull Request resolved: #124479 Approved by: https://github.com/jgong5, https://github.com/gujinghui, https://github.com/EikanWang, https://github.com/albanD ghstack dependencies: #124359

# Motivation This PR aims to refactor autocast **C++** APIs to be device-agnostic and deprecate the device-specific autocast **C++** APIs. In C++ side, - `is_enabled()` -> `is_enabled(device_type)`. - `set_enabled(new_enabled)` -> `set_enabled(device_type, new_enabled)`. - `get_autocast_dtype()` -> `get_autocast_dtype(device_type)` - `set_autocast_dtype(dtype)` -> `set_autocast_dtype(device_type, dtype)` These following C++ APIs are deprecated and should be removed in PyTorch 2.5 - `is_cpu_enabled` - `set_cpu_enabled` - `get_autocast_cpu_dtype` - `set_autocast_cpu_dtype` - `is_xpu_enabled` - `set_xpu_enabled` - `get_autocast_xpu_dtype` - `set_autocast_xpu_dtype` - `is_ipu_enabled` - `set_ipu_enabled` - `get_autocast_ipu_dtype` - `set_autocast_ipu_dtype` - `is_hpu_enabled` - `set_hpu_enabled` - `get_autocast_hpu_dtype` - `set_autocast_hpu_dtype` - `is_xla_enabled` - `set_xla_enabled` - `get_autocast_xla_dtype` - `set_autocast_xla_dtype` - `is_privateuseone_enabled` - `set_privateuseone_enabled` - `get_autocast_privateuseone_dtype` - `set_autocast_privateuseone_dtype` In Python side, provide 4 generic autocast APIs: - `torch.is_autocast_enabled(device_type)` - `torch.set_autocast_enabled(device_type, new_enabled)` - `torch.get_autocast_dtype(device_type)` - `torch.set_autocast_dtype(device_type, dtype)` # Additional Context We will submit another PR to refactor autocast **Python** APIs based on this PR. Pull Request resolved: #124359 Approved by: https://github.com/jgong5, https://github.com/albanD

# Motivation Refactor autocast usage scenario in `torch/amp/autocast_mode.py` and `torch/utils/checkpoint.py` to fix the bug - convention conflict between `torch.xxx.get_autocast_xxx_dtype` defined in `autocast_mode.py` and `torch.xxx.get_autocast_dtype` defined in `checkpoint.py`. # Solution Use device-agnostic APIs like `torch.get_autocast_dtype`, ..., instead. Pull Request resolved: #124479 Approved by: https://github.com/jgong5, https://github.com/gujinghui, https://github.com/EikanWang, https://github.com/albanD ghstack dependencies: #124359

# Motivation This PR aims to refactor autocast **C++** APIs to be device-agnostic and deprecate the device-specific autocast **C++** APIs. In C++ side, - `is_enabled()` -> `is_enabled(device_type)`. - `set_enabled(new_enabled)` -> `set_enabled(device_type, new_enabled)`. - `get_autocast_dtype()` -> `get_autocast_dtype(device_type)` - `set_autocast_dtype(dtype)` -> `set_autocast_dtype(device_type, dtype)` These following C++ APIs are deprecated and should be removed in PyTorch 2.5 - `is_cpu_enabled` - `set_cpu_enabled` - `get_autocast_cpu_dtype` - `set_autocast_cpu_dtype` - `is_xpu_enabled` - `set_xpu_enabled` - `get_autocast_xpu_dtype` - `set_autocast_xpu_dtype` - `is_ipu_enabled` - `set_ipu_enabled` - `get_autocast_ipu_dtype` - `set_autocast_ipu_dtype` - `is_hpu_enabled` - `set_hpu_enabled` - `get_autocast_hpu_dtype` - `set_autocast_hpu_dtype` - `is_xla_enabled` - `set_xla_enabled` - `get_autocast_xla_dtype` - `set_autocast_xla_dtype` - `is_privateuseone_enabled` - `set_privateuseone_enabled` - `get_autocast_privateuseone_dtype` - `set_autocast_privateuseone_dtype` In Python side, provide 4 generic autocast APIs: - `torch.is_autocast_enabled(device_type)` - `torch.set_autocast_enabled(device_type, new_enabled)` - `torch.get_autocast_dtype(device_type)` - `torch.set_autocast_dtype(device_type, dtype)` # Additional Context We will submit another PR to refactor autocast **Python** APIs based on this PR. Pull Request resolved: #124359 Approved by: https://github.com/jgong5, https://github.com/albanD

Refactor autocast usage scenario in `torch/amp/autocast_mode.py` and `torch/utils/checkpoint.py` to fix the bug - convention conflict between `torch.xxx.get_autocast_xxx_dtype` defined in `autocast_mode.py` and `torch.xxx.get_autocast_dtype` defined in `checkpoint.py`. Use device-agnostic APIs like `torch.get_autocast_dtype`, ..., instead. Pull Request resolved: #124479 Approved by: https://github.com/jgong5, https://github.com/gujinghui, https://github.com/EikanWang, https://github.com/albanD ghstack dependencies: #124359

This reverts commit 83106b7.

* Fix ref leak in `dtype.to_complex()`/`to_real()` (#125154) By using `Py_NewRef` Also, wrap `THPDtype_to_real`/`THPDtype_to_complex` calls with `HANDLE_TH_ERRORS` Add regression test for the above issues, by calling to_complex for integral dtypes, that raises an exception and by preserving reference count to the same to_complex/to_real call to detect if leak is happeneing. Replace ```cpp auto dtype = (PyObject*)torch::getTHPDtype(current_dtype); Py_INCREF(dtype); return dtype; ``` with a more compact/streamlined equivalent ```cpp return Py_NewRef(torch::getTHPDtype(current_dtype)); ``` Fixes #124868 Pull Request resolved: #125154 Approved by: https://github.com/Skylion007, https://github.com/albanD (cherry picked from commit 744f341) * Revert "Fix ref leak in `dtype.to_complex()`/`to_real()` (#125154)" This reverts commit a1b04d8. * Fix ref leak in `dtype.to_complex()`/`to_real()` (#125154) By using `Py_NewRef` Also, wrap `THPDtype_to_real`/`THPDtype_to_complex` calls with `HANDLE_TH_ERRORS` Add regression test for the above issues, by calling to_complex for integral dtypes, that raises an exception and by preserving reference count to the same to_complex/to_real call to detect if leak is happeneing. Replace ```cpp auto dtype = (PyObject*)torch::getTHPDtype(current_dtype); Py_INCREF(dtype); return dtype; ``` with a more compact/streamlined equivalent ```cpp return Py_NewRef(torch::getTHPDtype(current_dtype)); ``` Fixes #124868 Pull Request resolved: #125154 Approved by: https://github.com/Skylion007, https://github.com/albanD (cherry picked from commit 744f341) * Revert "Fix ref leak in `dtype.to_complex()`/`to_real()` (#125154)" This reverts commit 5a28bad. * Refactor autocast C++ APIs to be device-agnostic (#124359) # Motivation This PR aims to refactor autocast **C++** APIs to be device-agnostic and deprecate the device-specific autocast **C++** APIs. In C++ side, - `is_enabled()` -> `is_enabled(device_type)`. - `set_enabled(new_enabled)` -> `set_enabled(device_type, new_enabled)`. - `get_autocast_dtype()` -> `get_autocast_dtype(device_type)` - `set_autocast_dtype(dtype)` -> `set_autocast_dtype(device_type, dtype)` These following C++ APIs are deprecated and should be removed in PyTorch 2.5 - `is_cpu_enabled` - `set_cpu_enabled` - `get_autocast_cpu_dtype` - `set_autocast_cpu_dtype` - `is_xpu_enabled` - `set_xpu_enabled` - `get_autocast_xpu_dtype` - `set_autocast_xpu_dtype` - `is_ipu_enabled` - `set_ipu_enabled` - `get_autocast_ipu_dtype` - `set_autocast_ipu_dtype` - `is_hpu_enabled` - `set_hpu_enabled` - `get_autocast_hpu_dtype` - `set_autocast_hpu_dtype` - `is_xla_enabled` - `set_xla_enabled` - `get_autocast_xla_dtype` - `set_autocast_xla_dtype` - `is_privateuseone_enabled` - `set_privateuseone_enabled` - `get_autocast_privateuseone_dtype` - `set_autocast_privateuseone_dtype` In Python side, provide 4 generic autocast APIs: - `torch.is_autocast_enabled(device_type)` - `torch.set_autocast_enabled(device_type, new_enabled)` - `torch.get_autocast_dtype(device_type)` - `torch.set_autocast_dtype(device_type, dtype)` # Additional Context We will submit another PR to refactor autocast **Python** APIs based on this PR. Pull Request resolved: #124359 Approved by: https://github.com/jgong5, https://github.com/albanD * refactor autocast python APIs (#124479) Refactor autocast usage scenario in `torch/amp/autocast_mode.py` and `torch/utils/checkpoint.py` to fix the bug - convention conflict between `torch.xxx.get_autocast_xxx_dtype` defined in `autocast_mode.py` and `torch.xxx.get_autocast_dtype` defined in `checkpoint.py`. Use device-agnostic APIs like `torch.get_autocast_dtype`, ..., instead. Pull Request resolved: #124479 Approved by: https://github.com/jgong5, https://github.com/gujinghui, https://github.com/EikanWang, https://github.com/albanD ghstack dependencies: #124359 * Fix ref leak in `dtype.to_complex()`/`to_real()` (#125154) By using `Py_NewRef` Also, wrap `THPDtype_to_real`/`THPDtype_to_complex` calls with `HANDLE_TH_ERRORS` Add regression test for the above issues, by calling to_complex for integral dtypes, that raises an exception and by preserving reference count to the same to_complex/to_real call to detect if leak is happeneing. Replace ```cpp auto dtype = (PyObject*)torch::getTHPDtype(current_dtype); Py_INCREF(dtype); return dtype; ``` with a more compact/streamlined equivalent ```cpp return Py_NewRef(torch::getTHPDtype(current_dtype)); ``` Fixes #124868 Pull Request resolved: #125154 Approved by: https://github.com/Skylion007, https://github.com/albanD * Revert "refactor autocast python APIs (#124479)" This reverts commit 495b0c9. * Revert "Refactor autocast C++ APIs to be device-agnostic (#124359)" This reverts commit 83106b7. --------- Co-authored-by: Nikita Shulga <[email protected]> Co-authored-by: Huy Do <[email protected]> Co-authored-by: Yu, Guangye <[email protected]>

guangyey requested review from albanD and soulitzer as code owners April 18, 2024 05:13

guangyey mentioned this pull request Apr 18, 2024

Support xpu autocast policy #124052

Closed

pytorch-bot bot added the module: amp (automated mixed precision) autocast label Apr 18, 2024

guangyey marked this pull request as draft April 18, 2024 05:15

guangyey changed the title ~~refactor autocast APIs to be device-agnostic~~ [WIP] Refactor autocast APIs to be device-agnostic Apr 18, 2024

pytorchbot added the open source label Apr 18, 2024

guangyey added a commit that referenced this pull request Apr 18, 2024

refactor autocast APIs to be device-agnostic

b46c0f7

ghstack-source-id: eeb80ac Pull Request resolved: #124359

pytorch-bot bot added the release notes: jit release notes category label Apr 18, 2024

guangyey added 4 commits April 18, 2024 13:01

refactor autocast APIs to be device-agnostic

8b256b0

[ghstack-poisoned]

Update on "[WIP] Refactor autocast APIs to be device-agnostic"

6533924

cc mcarilli ptrblck leslie-fang-intel jgong5 [ghstack-poisoned]

Update on "[WIP] Refactor autocast APIs to be device-agnostic"

eb26f91

cc mcarilli ptrblck leslie-fang-intel jgong5 [ghstack-poisoned]

Update on "[WIP] Refactor autocast APIs to be device-agnostic"

0dd786e

cc mcarilli ptrblck leslie-fang-intel jgong5 [ghstack-poisoned]

guangyey added 2 commits April 18, 2024 17:50

Update on "[WIP] Refactor autocast APIs to be device-agnostic"

86a8066

cc mcarilli ptrblck leslie-fang-intel jgong5 [ghstack-poisoned]

albanD approved these changes Apr 22, 2024

View reviewed changes

albanD mentioned this pull request Apr 22, 2024

is_any_autocast_enabled support privateuseone. #124585

Closed

bhack mentioned this pull request Apr 22, 2024

Force upsample to be float32 #121324

Closed

guangyey commented Apr 23, 2024

View reviewed changes

guangyey added ciflow/periodic Trigger jobs ran periodically on master (periodic.yml) on the PR ciflow/rocm Trigger "default" config CI on ROCm and removed ciflow/rocm Trigger "default" config CI on ROCm labels Apr 23, 2024

pytorchmergebot added the merging label Apr 23, 2024

guangyey added ciflow/rocm Trigger "default" config CI on ROCm and removed ciflow/rocm Trigger "default" config CI on ROCm labels Apr 23, 2024

pytorchmergebot added the Merged label Apr 23, 2024

pytorchmergebot closed this in 25f321b Apr 23, 2024

pytorchmergebot removed the merging label Apr 23, 2024

guangyey mentioned this pull request Apr 25, 2024

add new API torch.amp.is_autocast_available #124938

Closed

carmocca mentioned this pull request May 13, 2024

Avoid deprecated autocast APIs Lightning-AI/lightning-thunder#412

Closed

huydhn added a commit that referenced this pull request May 14, 2024

Revert "Refactor autocast C++ APIs to be device-agnostic (#124359)"

730b4dd

This reverts commit 83106b7.

github-actions bot deleted the gh/guangyey/23/head branch June 2, 2024 02:04

FFFrog mentioned this pull request Aug 14, 2024

Keep up to date cosdt/torch_backend#81

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Refactor autocast C++ APIs to be device-agnostic #124359

Refactor autocast C++ APIs to be device-agnostic #124359

Uh oh!

guangyey commented Apr 18, 2024 •

edited by pytorch-bot bot

Loading

Uh oh!

pytorch-bot bot commented Apr 18, 2024 •

edited

Loading

Uh oh!

guangyey commented Apr 18, 2024 •

edited

Loading

Uh oh!

albanD commented Apr 18, 2024

Uh oh!

guangyey commented Apr 19, 2024

Uh oh!

guangyey Apr 23, 2024

Uh oh!

guangyey commented Apr 23, 2024

Uh oh!

pytorchmergebot commented Apr 23, 2024

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Refactor autocast C++ APIs to be device-agnostic #124359

Refactor autocast C++ APIs to be device-agnostic #124359

Uh oh!

Conversation

guangyey commented Apr 18, 2024 • edited by pytorch-bot bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Motivation

Additional Context

Uh oh!

pytorch-bot bot commented Apr 18, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/124359

✅ You can merge normally! (1 Unrelated Failure)

Uh oh!

guangyey commented Apr 18, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

albanD commented Apr 18, 2024

Uh oh!

guangyey commented Apr 19, 2024

Uh oh!

guangyey Apr 23, 2024

Choose a reason for hiding this comment

Uh oh!

guangyey commented Apr 23, 2024

Uh oh!

pytorchmergebot commented Apr 23, 2024

Merge started

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

guangyey commented Apr 18, 2024 •

edited by pytorch-bot bot

Loading

pytorch-bot bot commented Apr 18, 2024 •

edited

Loading

guangyey commented Apr 18, 2024 •

edited

Loading