Aten _To_Copy #6055
Open
Aten _To_Copy #6055
+502
−139
Conversation
This file contains 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
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/6055
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New FailureAs of commit 5b4ff1f with merge base 27330f2 ( NEW FAILURE - The following job has failed:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D64080303 |
SS-JIA
approved these changes
Oct 9, 2024
|
This pull request was exported from Phabricator. Differential Revision: D64080303 |
Abhi-hpp
added a commit
to Abhi-hpp/executorch
that referenced
this pull request
Oct 9, 2024
Summary: Pull Request resolved: pytorch#6055 Implement aten._to_copy. Currently we are only interested in fp32 <-> fp16 conversions, but it should theoritically support other dtype conversions too. I noticed an issue with int conversions so limited it to just fp32 and fp16 for now. Note: Most driver implementations of fp16 cast does not "round up" the result, therefore there might be 1 bit difference between vulkan output and cpu torch.to. Explained in greater detail in the comments. Reviewed By: SS-JIA Differential Revision: D64080303
Abhi-hpp
force-pushed
the
export-D64080303
branch
from
c2297a5
to
8bd0580
Compare
|
This pull request was exported from Phabricator. Differential Revision: D64080303 |
Abhi-hpp
added a commit
to Abhi-hpp/executorch
that referenced
this pull request
Oct 9, 2024
Summary: Pull Request resolved: pytorch#6055 Implement aten._to_copy. Currently we are only interested in fp32 <-> fp16 conversions, but it should theoritically support other dtype conversions too. I noticed an issue with int conversions so limited it to just fp32 and fp16 for now. Note: Most driver implementations of fp16 cast does not "round up" the result, therefore there might be 1 bit difference between vulkan output and cpu torch.to. Explained in greater detail in the comments. Reviewed By: SS-JIA Differential Revision: D64080303
Abhi-hpp
force-pushed
the
export-D64080303
branch
from
8bd0580
to
858723c
Compare
|
This pull request was exported from Phabricator. Differential Revision: D64080303 |
Abhi-hpp
added a commit
to Abhi-hpp/executorch
that referenced
this pull request
Oct 9, 2024
Summary: Pull Request resolved: pytorch#6055 Implement aten._to_copy. Currently we are only interested in fp32 <-> fp16 conversions, but it should theoritically support other dtype conversions too. I noticed an issue with int conversions so limited it to just fp32 and fp16 for now. Note: Most driver implementations of fp16 cast does not "round up" the result, therefore there might be 1 bit difference between vulkan output and cpu torch.to. Explained in greater detail in the comments. Reviewed By: SS-JIA Differential Revision: D64080303
Abhi-hpp
force-pushed
the
export-D64080303
branch
from
858723c
to
7a0be72
Compare
|
This pull request was exported from Phabricator. Differential Revision: D64080303 |
Abhi-hpp
added a commit
to Abhi-hpp/executorch
that referenced
this pull request
Oct 9, 2024
Summary: Pull Request resolved: pytorch#6055 Implement aten._to_copy. Currently we are only interested in fp32 <-> fp16 conversions, but it should theoritically support other dtype conversions too. I noticed an issue with int conversions so limited it to just fp32 and fp16 for now. Note: Most driver implementations of fp16 cast does not "round up" the result, therefore there might be 1 bit difference between vulkan output and cpu torch.to. Explained in greater detail in the comments. Reviewed By: SS-JIA Differential Revision: D64080303
Abhi-hpp
force-pushed
the
export-D64080303
branch
from
7a0be72
to
adcc2f7
Compare
Abhi-hpp
added a commit
to Abhi-hpp/executorch
that referenced
this pull request
Oct 9, 2024
Summary: Pull Request resolved: pytorch#6055 Implement aten._to_copy. Currently we are only interested in fp32 <-> fp16 conversions, but it should theoritically support other dtype conversions too. I noticed an issue with int conversions so limited it to just fp32 and fp16 for now. Note: Most driver implementations of fp16 cast does not "round up" the result, therefore there might be 1 bit difference between vulkan output and cpu torch.to. Explained in greater detail in the comments. Reviewed By: SS-JIA Differential Revision: D64080303
Abhi-hpp
force-pushed
the
export-D64080303
branch
from
adcc2f7
to
2c3a809
Compare
|
This pull request was exported from Phabricator. Differential Revision: D64080303 |
added 2 commits
October 9, 2024 16:33
Summary: Introduce Graph node which will be the parent class for nodes. This allows us to have different nodes that correspond to different vk command buffer functions including shader dispatch. Differential Revision: D64080291
Abhi-hpp
added a commit
to Abhi-hpp/executorch
that referenced
this pull request
Oct 9, 2024
Summary: Pull Request resolved: pytorch#6055 Implement aten._to_copy. Currently we are only interested in fp32 <-> fp16 conversions, but it should theoritically support other dtype conversions too. I noticed an issue with int conversions so limited it to just fp32 and fp16 for now. Note: Most driver implementations of fp16 cast does not "round up" the result, therefore there might be 1 bit difference between vulkan output and cpu torch.to. Explained in greater detail in the comments. Reviewed By: SS-JIA Differential Revision: D64080303
Abhi-hpp
force-pushed
the
export-D64080303
branch
from
2c3a809
to
c369ff7
Compare
|
This pull request was exported from Phabricator. Differential Revision: D64080303 |
Summary: Pull Request resolved: pytorch#6055 Implement aten._to_copy. Currently we are only interested in fp32 <-> fp16 conversions, but it should theoritically support other dtype conversions too. I noticed an issue with int conversions so limited it to just fp32 and fp16 for now. Note: Most driver implementations of fp16 cast does not "round up" the result, therefore there might be 1 bit difference between vulkan output and cpu torch.to. Explained in greater detail in the comments. Reviewed By: SS-JIA Differential Revision: D64080303
|
This pull request was exported from Phabricator. Differential Revision: D64080303 |
Abhi-hpp
force-pushed
the
export-D64080303
branch
from
c369ff7
to
5b4ff1f
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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:
Implement aten._to_copy. Currently we are only interested in fp32 <-> fp16 conversions, but it should theoritically support other dtype conversions too. I noticed an issue with int conversions so limited it to just fp32 and fp16 for now.
Note: Most driver implementations of fp16 cast does not "round up" the result, therefore there might be 1 bit difference between vulkan output and cpu torch.to. Explained in greater detail in the comments.
Differential Revision: D64080303