[quant][core][gpu][feature] Implemented quantized cuda gelu by dzdang · Pull Request #77212 · pytorch/pytorch

dzdang · 2022-05-11T00:53:14Z

Stack from ghstack (oldest at bottom):

-> [quant][core][gpu][feature] Implemented quantized cuda gelu #77212

Summary:
Support for quantized cuda gelu has been provided by using
dequantize -> fp32 cuda gelu kernel -> quantize. Mathematically, this
is not equivalent to doing int8 gelu, so we have opted for this approach
for now. It might be possible to write a variant of the int8 gelu that's
equivalent to dequantize -> fp32 cuda gelu kernel -> quantize, which
can be a topic for future work.

Test function test_qgelu was amended to test gelu for quantized cuda
backends.

Test Plan:

python test/test_quantization.py -k test_qgelu

Differential Revision: D36392774

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` [ghstack-poisoned]

facebook-github-bot · 2022-05-11T00:53:19Z

🔗 Helpful links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/77212
📄 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 our office hours
↩️ [fb-only] Re-run with SSH instructions

❌ 2 New Failures

As of commit a8d5169 (more details on the Dr. CI page):

Expand to see more

2/2 failures introduced in this PR

🕵️ 2 new failures recognized by patterns

The following CI failures do not appear to be due to upstream breakages

pull / pytorch-xla-linux-bionic-py3.7-clang8 / test (xla, 1, 1, linux.2xlarge) (1/2)

Step: "Test" (full log | diagnosis details | 🔁 rerun)

2022-05-25T00:39:26.7645009Z AssertionError: Missing shape inference function.

2022-05-25T00:39:25.0487009Z PyTorch Commit ID: a8d516984b632f7d56699acaad6d67598f6ff419
2022-05-25T00:39:25.0555048Z /var/lib/jenkins/workspace /var/lib/jenkins/workspace/xla
2022-05-25T00:39:26.0951135Z /var/lib/jenkins/workspace/xla
2022-05-25T00:39:26.7641584Z Traceback (most recent call last):
2022-05-25T00:39:26.7642243Z   File "/var/lib/jenkins/workspace/xla/scripts/gen_lazy_tensor.py", line 83, in <module>
2022-05-25T00:39:26.7642855Z     get_device_fn="torch_xla::bridge::GetXlaDevice")
2022-05-25T00:39:26.7643491Z   File "/opt/conda/lib/python3.7/site-packages/torchgen/gen_lazy_tensor.py", line 409, in run_gen_lazy_tensor
2022-05-25T00:39:26.7643851Z     validate_shape_inference_header(shape_inference_hdr, expected_shape_infr_decls)
2022-05-25T00:39:26.7644315Z   File "/opt/conda/lib/python3.7/site-packages/torchgen/gen_lazy_tensor.py", line 161, in validate_shape_inference_header
2022-05-25T00:39:26.7644597Z     {decl}"""
2022-05-25T00:39:26.7645009Z AssertionError: Missing shape inference function.
2022-05-25T00:39:26.7645165Z 
2022-05-25T00:39:26.7645347Z Please add declare this function in /var/lib/jenkins/workspace/torch/csrc/lazy/core/shape_inference.h:
2022-05-25T00:39:26.7645556Z 
2022-05-25T00:39:26.7645700Z and implement it in the the corresponding shape_inference.cpp file.
2022-05-25T00:39:26.7645871Z 
2022-05-25T00:39:26.7646020Z TORCH_API std::vector<torch::lazy::Shape> compute_shape_inverse(const at::Tensor & self);
2022-05-25T00:39:26.8195127Z Failed to generate lazy files: ['python', '/var/lib/jenkins/workspace/xla/scripts/gen_lazy_tensor.py']
2022-05-25T00:39:26.9652218Z + cleanup
2022-05-25T00:39:26.9652526Z + retcode=1
2022-05-25T00:39:26.9653014Z + set +x

pull / linux-xenial-py3.7-gcc5.4 / test (backwards_compat, 1, 1, linux.2xlarge) (2/2)

Step: "Test" (full log | diagnosis details | 🔁 rerun)

2022-05-25T00:33:37.0188936Z The PR is introduc...m to confirm whether this change is wanted or not.

2022-05-25T00:33:37.0175536Z processing existing schema:  text(__torch__.torch.classes.profiling.SourceRef _0) -> (str _0)
2022-05-25T00:33:37.0177019Z processing existing schema:  count(__torch__.torch.classes.profiling.InstructionStats _0) -> (int _0)
2022-05-25T00:33:37.0177926Z processing existing schema:  duration_ns(__torch__.torch.classes.profiling.InstructionStats _0) -> (int _0)
2022-05-25T00:33:37.0179458Z processing existing schema:  source(__torch__.torch.classes.profiling.SourceStats _0) -> (__torch__.torch.classes.profiling.SourceRef _0)
2022-05-25T00:33:37.0181278Z processing existing schema:  line_map(__torch__.torch.classes.profiling.SourceStats _0) -> (Dict(int, __torch__.torch.classes.profiling.InstructionStats) _0)
2022-05-25T00:33:37.0182081Z processing existing schema:  __init__(__torch__.torch.classes.profiling._ScriptProfile _0) -> (NoneType _0)
2022-05-25T00:33:37.0183371Z processing existing schema:  enable(__torch__.torch.classes.profiling._ScriptProfile _0) -> (NoneType _0)
2022-05-25T00:33:37.0184430Z processing existing schema:  disable(__torch__.torch.classes.profiling._ScriptProfile _0) -> (NoneType _0)
2022-05-25T00:33:37.0187271Z processing existing schema:  _dump_stats(__torch__.torch.classes.profiling._ScriptProfile _0) -> (__torch__.torch.classes.profiling.SourceStats[] _0)
2022-05-25T00:33:37.0188127Z processing existing schema:  __init__(__torch__.torch.classes.dist_rpc.WorkerInfo _0, str _1, int _2) -> (NoneType _0)
2022-05-25T00:33:37.0188936Z The PR is introducing backward incompatible changes to the operator library. Please contact PyTorch team to confirm whether this change is wanted or not. 
2022-05-25T00:33:37.0188948Z 
2022-05-25T00:33:37.0189066Z Broken ops: [
2022-05-25T00:33:37.0189642Z 	prims::uniform(int[] shape, *, Scalar low, Scalar high, int dtype, Device device) -> (Tensor)
2022-05-25T00:33:37.0192176Z 	prims::full_like(Tensor a, Scalar fill_value, *, int dtype, Device device, bool requires_grad) -> (Tensor)
2022-05-25T00:33:37.0192662Z 	prims::full(int[] shape, Scalar fill_value, *, int dtype, Device device, bool requires_grad) -> (Tensor)
2022-05-25T00:33:37.0193174Z 	prims::empty_strided(int[] shape, int[] strides, *, int dtype, Device device, bool requires_grad) -> (Tensor)
2022-05-25T00:33:37.0194244Z 	prims::empty(int[] shape, *, int dtype, Device device, bool requires_grad) -> (Tensor)
2022-05-25T00:33:37.0194604Z 	prims::amin(Tensor inp, int[]? dims, *, int? output_dtype=None) -> (Tensor)
2022-05-25T00:33:37.0194947Z 	prims::amax(Tensor inp, int[]? dims, *, int? output_dtype=None) -> (Tensor)
2022-05-25T00:33:37.0195370Z 	prims::var(Tensor inp, int[]? dims, *, int correction, int? output_dtype=None) -> (Tensor)

This comment was automatically generated by Dr. CI (expand for details).

Please report bugs/suggestions to the (internal) Dr. CI Users group.

Click here to manually regenerate this comment.

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` ghstack-source-id: 7f74cbc Pull Request resolved: #77212

dzdang · 2022-05-11T00:54:08Z

@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` Differential Revision: [D36302475](https://our.internmc.facebook.com/intern/diff/D36302475) [ghstack-poisoned]

dzdang · 2022-05-11T16:34:09Z

@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` Differential Revision: [D36302475](https://our.internmc.facebook.com/intern/diff/D36302475) [ghstack-poisoned]

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` ghstack-source-id: 7ac334a Pull Request resolved: #77212

dzdang · 2022-05-11T16:49:07Z

@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

dzdang · 2022-05-11T17:03:07Z

@jerryzh168 wondering if we should move Activation.cpp in quantized/cuda instead since currently it has nothing to do with cuDNN?

jerryzh168 · 2022-05-11T17:07:13Z

yeah I think maybe we can move it to cuda folder

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` Differential Revision: [D36302475](https://our.internmc.facebook.com/intern/diff/D36302475) [ghstack-poisoned]

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` ghstack-source-id: 2bec24f Pull Request resolved: #77212

dzdang · 2022-05-11T18:41:36Z

@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

jerryzh168 · 2022-05-12T17:28:24Z

aten/src/ATen/native/quantized/cuda/Activation.cpp

+  auto x_fp32 = at::dequantize(qx);
+  auto result_fp32 = at::gelu(x_fp32);
+  return at::quantize_per_tensor(result_fp32, qx.q_scale(), qx.q_zero_point(), qx.scalar_type());


if each one of them supports L11-13 we can remove L11-13 right? should we add a TODO to do this?

they do, but I think I've seen several other functions this use pruning check for early termination

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` Differential Revision: [D36302475](https://our.internmc.facebook.com/intern/diff/D36302475) [ghstack-poisoned]

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` ghstack-source-id: 0d104e9 Pull Request resolved: #77212

dzdang · 2022-05-13T13:07:24Z

@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

facebook-github-bot · 2022-05-13T20:58:06Z

@pytorchbot merge this

(Initiating merge automatically since Phabricator Diff has merged)

Summary: Pull Request resolved: #77212 Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` Reviewed By: jerryzh168 Differential Revision: D36302475 Pulled By: dzdang fbshipit-source-id: 11342fb290031d62ba5e620cbe572fe2cc8ed701

malfet · 2022-05-14T00:13:45Z

Reverting this PR internally as it broke bazel builds, see https://hud.pytorch.org/pytorch/pytorch/commit/b892b85b881c7b3b2b6bde529c4d174e348ba9fb

facebook-github-bot · 2022-05-14T00:16:25Z

@pytorchbot revert this

This Pull Request has been reverted by a revert inside Meta. To re-land this change, please open another pull request, assign the same reviewers, fix the CI failures that caused the revert and make sure that the failing CI runs on the PR by applying the proper ciflow label (e.g., ciflow/trunk).)

This reverts commit b892b85. Reverted #77212 on behalf of https://github.com/facebook-github-bot

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` Differential Revision: [D36302475](https://our.internmc.facebook.com/intern/diff/D36302475) [ghstack-poisoned]

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` ghstack-source-id: e617725 Pull Request resolved: #77212

dzdang · 2022-05-14T00:40:27Z

@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` Differential Revision: [D36392774](https://our.internmc.facebook.com/intern/diff/D36392774) [ghstack-poisoned]

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` ghstack-source-id: 31203eb Pull Request resolved: #77212

dzdang · 2022-05-17T13:06:51Z

@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` Differential Revision: [D36392774](https://our.internmc.facebook.com/intern/diff/D36392774) [ghstack-poisoned]

Summary: Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` ghstack-source-id: 5b2a79f Pull Request resolved: #77212

dzdang · 2022-05-17T18:23:37Z

@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

facebook-github-bot · 2022-05-24T22:30:11Z

@pytorchbot merge

(Initiating merge automatically since Phabricator Diff has merged)

facebook-github-bot · 2022-05-24T22:30:17Z

@pytorchbot merge

(Initiating merge automatically since Phabricator Diff has merged)

…36302475) (#77212) Summary: Pull Request resolved: #77212 Support for quantized cuda gelu has been provided by using `dequantize -> fp32 cuda gelu kernel -> quantize`. Mathematically, this is not equivalent to doing int8 gelu, so we have opted for this approach for now. It might be possible to write a variant of the int8 gelu that's equivalent to `dequantize -> fp32 cuda gelu kernel -> quantize`, which can be a topic for future work. Test function `test_qgelu` was amended to test gelu for quantized cuda backends. Test Plan: ``` python test/test_quantization.py -k test_qgelu ``` Reviewed By: cpuhrsch Differential Revision: D36392774 Pulled By: dzdang fbshipit-source-id: 1accdefb042ee4930451ef016c527c5cd3e13168

pytorchmergebot · 2022-05-24T22:33:51Z

Can't merge closed PR #77212

dzdang · 2022-05-24T23:03:25Z

@pytorchbot merge

pytorchmergebot · 2022-05-24T23:04:36Z

Can't merge closed PR #77212

dzdang · 2022-05-25T00:18:39Z

@pytorchbot merge

pytorchmergebot · 2022-05-25T00:21:30Z

dzdang · 2022-05-25T13:13:58Z

@pytorchbot merge

pytorchmergebot · 2022-05-25T13:15:13Z

Merge failed due to Command git -C /home/runner/actions-runner/_work/pytorch/pytorch cherry-pick -x 49eb6f09a999640d3c2ccebfcd11c10d7c16f2d2 returned non-zero exit code 1

Auto-merging BUILD.bazel
Auto-merging aten/src/ATen/native/native_functions.yaml
The previous cherry-pick is now empty, possibly due to conflict resolution.
If you wish to commit it anyway, use:

    git commit --allow-empty

Otherwise, please use 'git cherry-pick --skip'
On branch master
Your branch is up to date with 'origin/master'.

You are currently cherry-picking commit 49eb6f09a9.
  (all conflicts fixed: run "git cherry-pick --continue")
  (use "git cherry-pick --skip" to skip this patch)
  (use "git cherry-pick --abort" to cancel the cherry-pick operation)

nothing to commit, working tree clean

Raised by https://github.com/pytorch/pytorch/actions/runs/2384646215

dzdang requested a review from bdhirsh as a code owner May 11, 2022 00:53

facebook-github-bot added the cla signed label May 11, 2022

dzdang added release notes: quantization release notes category topic: new features topic category labels May 11, 2022

dzdang added 2 commits May 11, 2022 09:38

dzdang requested a review from jerryzh168 May 11, 2022 17:01

jerryzh168 reviewed May 12, 2022

View reviewed changes

jerryzh168 approved these changes May 12, 2022

View reviewed changes

pytorchmergebot added the Merged label May 13, 2022

pytorchmergebot closed this in b892b85 May 13, 2022

pytorchmergebot added the Reverted label May 14, 2022

pytorchmergebot added a commit that referenced this pull request May 14, 2022

Revert "[quant][core][gpu][feature] Implemented quantized cuda gelu"

981719f

This reverts commit b892b85. Reverted #77212 on behalf of https://github.com/facebook-github-bot

dzdang reopened this May 14, 2022

janeyx99 mentioned this pull request May 18, 2022

[Meta] CI Revert Tracker #66178

Closed

pytorchmergebot closed this in 2aad28a May 24, 2022

dzdang reopened this May 25, 2022

dzdang closed this May 25, 2022

facebook-github-bot deleted the gh/dzdang/110/head branch May 28, 2022 14:17

Conversation

dzdang commented May 11, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

facebook-github-bot commented May 11, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful links

❌ 2 New Failures

🕵️ 2 new failures recognized by patterns

pull / pytorch-xla-linux-bionic-py3.7-clang8 / test (xla, 1, 1, linux.2xlarge) (1/2)

pull / linux-xenial-py3.7-gcc5.4 / test (backwards_compat, 1, 1, linux.2xlarge) (2/2)

Uh oh!

dzdang commented May 11, 2022

Uh oh!

dzdang commented May 11, 2022

Uh oh!

dzdang commented May 11, 2022

Uh oh!

dzdang commented May 11, 2022

Uh oh!

jerryzh168 commented May 11, 2022

Uh oh!

dzdang commented May 11, 2022

Uh oh!

jerryzh168 May 12, 2022

Choose a reason for hiding this comment

Uh oh!

dzdang May 12, 2022

Choose a reason for hiding this comment

Uh oh!

dzdang commented May 13, 2022

Uh oh!

facebook-github-bot commented May 13, 2022

Uh oh!

malfet commented May 14, 2022

Uh oh!

facebook-github-bot commented May 14, 2022

Uh oh!

dzdang commented May 14, 2022

Uh oh!

dzdang commented May 17, 2022

Uh oh!

dzdang commented May 17, 2022

Uh oh!

facebook-github-bot commented May 24, 2022

Uh oh!

facebook-github-bot commented May 24, 2022

Uh oh!

pytorchmergebot commented May 24, 2022

Uh oh!

dzdang commented May 24, 2022

Uh oh!

pytorchmergebot commented May 24, 2022

Uh oh!

dzdang commented May 25, 2022

Uh oh!

pytorchmergebot commented May 25, 2022

Uh oh!

dzdang commented May 25, 2022

Uh oh!

pytorchmergebot commented May 25, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

dzdang commented May 11, 2022 •

edited

Loading

facebook-github-bot commented May 11, 2022 •

edited

Loading