New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Use c10 version of half/bfloat16 in executorch #7040

Open

swolchok wants to merge 38 commits into gh/swolchok/119/base from gh/swolchok/119/head

Contributor

swolchok commented Nov 23, 2024 •

edited

Loading

Stack from ghstack (oldest at bottom):

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using using in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: D66106969


          Use c10 version of half/bfloat16 in executorch

906f932

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

pytorch-bot bot commented Nov 23, 2024 •

edited

Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/7040

📄 Preview Python docs built from this PR

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

❌ 3 New Failures, 4 Cancelled Jobs, 1 Unrelated Failure

As of commit 8b8987a with merge base c28ae4a ():

NEW FAILURES - The following jobs have failed:

Apple / build-benchmark-app / macos-job (gh)
RuntimeError: Command bash /Users/runner/work/_temp/exec_script failed with exit code 1
pull / unittest / linux / linux-job (gh)
RuntimeError: Command docker exec -t e95738d4950a23609ff96f3237bee628903f198ad063abbed214d8ed531ac5bc /exec failed with exit code 8
trunk / test-models-macos (cmake, emformer_transcribe, xnnpack-quantization-delegation, macos-m1-stable, 90) / macos-job (gh)
RuntimeError: Missing out variants: {'quantized_decomposed::dequantize_per_tensor', 'quantized_decomposed::quantize_per_tensor'}

CANCELLED JOBS - The following jobs were cancelled. Please retry:

pull / test-eval_llama-mmlu-linux / linux-job (gh)
##[error]The operation was canceled.
pull / test-llava-runner-linux / linux-job (gh)
##[error]The operation was canceled.
trunk / test-models-macos (cmake, add, xnnpack-quantization-delegation, macos-m1-stable, 90) / macos-job (gh)
##[error]The operation was canceled.
trunk / test-models-macos (cmake, mv2, xnnpack-quantization-delegation, macos-m1-stable, 90) / macos-job (gh)
##[error]The operation was canceled.

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

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

pull / unittest / macos / macos-job (gh) (trunk failure)
backends/xnnpack/test/ops/test_conv1d.py::TestConv1d::test_qs8_conv1d_batchnorm_seq

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

facebook-github-bot added the CLA Signed label

Contributor

facebook-github-bot commented Nov 23, 2024

This pull request was exported from Phabricator. Differential Revision: D66106969

facebook-github-bot added the fb-exported label

This was referenced Nov 23, 2024

Reuse GELU implementation from PyTorch core #7041

Open

Remove ExecuTorch copy of Vectorized #7042

Open

Contributor Author

swolchok commented Nov 23, 2024

Working on fixing the build failures. This is going to require a bump of the PyTorch nightly to include pytorch/pytorch@f630799 (landed November 18).

swolchok added the topic: not user facing label


          Update on "Use c10 version of half/bfloat16 in executorch"

4b91ca9

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Nov 23, 2024

This pull request was exported from Phabricator. Differential Revision: D66106969

swolchok added a commit that referenced this pull request


          Bump ExecuTorch's PyTorch nightly pin to dev20241121

cdf0625

Require at least 11/18 to unblock #7040 .

Differential Revision: [D66398425](https://our.internmc.facebook.com/intern/diff/D66398425/)

[ghstack-poisoned]

swolchok mentioned this pull request

Bump ExecuTorch's PyTorch nightly pin to dev20241121 #7044

Closed


          Update on "Use c10 version of half/bfloat16 in executorch"

bb52821

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Nov 23, 2024

This pull request was exported from Phabricator. Differential Revision: D66106969


          Update on "Use c10 version of half/bfloat16 in executorch"

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Nov 26, 2024

This pull request was exported from Phabricator. Differential Revision: D66106969

swolchok added a commit that referenced this pull request


          Update base for Update on "Bump ExecuTorch's PyTorch nightly pin to d…

8c15b6c

…ev20241121"

Require at least 11/18 to unblock #7040 .

Differential Revision: [D66398425](https://our.internmc.facebook.com/intern/diff/D66398425/)

[ghstack-poisoned]

swolchok added a commit that referenced this pull request


          Update on "Bump ExecuTorch's PyTorch nightly pin to dev20241121"

aaa7768

Require at least 11/18 to unblock #7040 .

Differential Revision: [D66398425](https://our.internmc.facebook.com/intern/diff/D66398425/)

[ghstack-poisoned]


          Update on "Use c10 version of half/bfloat16 in executorch"

44dd36f

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Dec 2, 2024

This pull request was exported from Phabricator. Differential Revision: D66106969

swolchok added a commit that referenced this pull request


          Update base for Update on "Bump ExecuTorch's PyTorch nightly pin to d…

71801b1

…ev20241121"

Require at least 11/18 to unblock #7040 .

Differential Revision: [D66398425](https://our.internmc.facebook.com/intern/diff/D66398425/)

[ghstack-poisoned]

swolchok added a commit that referenced this pull request


          Update on "Bump ExecuTorch's PyTorch nightly pin to dev20241121"

2ed5b98

Require at least 11/18 to unblock #7040 .

Differential Revision: [D66398425](https://our.internmc.facebook.com/intern/diff/D66398425/)

[ghstack-poisoned]


          Update on "Use c10 version of half/bfloat16 in executorch"

4975faf

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Dec 2, 2024

This pull request was exported from Phabricator. Differential Revision: D66106969

swolchok added a commit that referenced this pull request


          Update base for Update on "Bump ExecuTorch's PyTorch nightly pin to d…

ec7367c

…ev20241121"

Require at least 11/18 to unblock #7040 .

Differential Revision: [D66398425](https://our.internmc.facebook.com/intern/diff/D66398425/)

[ghstack-poisoned]

swolchok added a commit that referenced this pull request


          Update on "Bump ExecuTorch's PyTorch nightly pin to dev20241121"

0520d25

Require at least 11/18 to unblock #7040 .

Differential Revision: [D66398425](https://our.internmc.facebook.com/intern/diff/D66398425/)

[ghstack-poisoned]


          rebase on "Use c10 version of half/bfloat16 in executorch"

a94c90a

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Dec 8, 2024

This pull request was exported from Phabricator. Differential Revision: D66106969

swolchok added a commit that referenced this pull request


          Update base for rebase on "Bump ExecuTorch's PyTorch nightly pin to d…

000e88f

…ev20241121"

Require at least 11/18 to unblock #7040 .

Differential Revision: [D66398425](https://our.internmc.facebook.com/intern/diff/D66398425/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Jan 9, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969

This was referenced Jan 9, 2025

stop double-installing ExecuTorch in one-off linux jobs #7578

Closed

use-pt-pinned-commit for test-arm-{backend,reference}-delegation #7579

Open

swolchok added the ciflow/trunk label


          Update on "Use c10 version of half/bfloat16 in executorch"

ef0e952

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Jan 10, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969


          Update on "Use c10 version of half/bfloat16 in executorch"

69e2672

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

swolchok pushed a commit that referenced this pull request


          Use c10 version of half/bfloat16 in executorch

62e2c38

Pull Request resolved: #7040

Pull Request resolved: pytorch/pytorch#144111

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.
ghstack-source-id: 260916624
@exported-using-ghexport

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

Contributor

facebook-github-bot commented Jan 10, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969


          Update on "Use c10 version of half/bfloat16 in executorch"

f4c5b77

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Jan 10, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969


          Update on "Use c10 version of half/bfloat16 in executorch"

ac4c80b

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Jan 10, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969


          Update on "Use c10 version of half/bfloat16 in executorch"

26c6211

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Jan 11, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969

shoumikhin approved these changes

View reviewed changes


          Update on "Use c10 version of half/bfloat16 in executorch"

8681d7f

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Jan 11, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969

swolchok pushed a commit that referenced this pull request


          Use c10 version of half/bfloat16 in executorch

975c5f2

Pull Request resolved: #7040

Pull Request resolved: pytorch/pytorch#144111

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.
ghstack-source-id: 261000207
@exported-using-ghexport

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)


          Update on "Use c10 version of half/bfloat16 in executorch"

bb69dcd

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Jan 13, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969


          Update on "Use c10 version of half/bfloat16 in executorch"

5a7cba3

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Jan 15, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969


          Update on "Use c10 version of half/bfloat16 in executorch"

abe03ea

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Jan 15, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969


          Update on "Use c10 version of half/bfloat16 in executorch"

789c521

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Jan 17, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969


          Update on "Use c10 version of half/bfloat16 in executorch"

8b8987a

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

[ghstack-poisoned]

Contributor

facebook-github-bot commented Jan 24, 2025

This pull request was exported from Phabricator. Differential Revision: D66106969

swolchok mentioned this pull request

--use-pt-pinned-commit in one-off linux jobs #7922

Open

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ciflow/trunk CLA Signed fb-exported topic: not user facing