Replace torch einsum with opt_einsum #440
Closed
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
facebook-github-bot
added
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D37128344 |
|
Optimized einsums obtain significant trends towards an effective reduction in runtime. The most impact is from convolution and linear layers uniformly for all layer sizes. This is under 10 iterations of benchmarking. Runtime_x is from Torch einsums and runtime_y is from opt einsums (in table below).
|
|
This pull request was exported from Phabricator. Differential Revision: D37128344 |
sayanghosh
added a commit
to sayanghosh/opacus
that referenced
this pull request
Jun 23, 2022
Summary: Pull Request resolved: pytorch#440 We are using optimized einsums in place of Pytorch einsums. As per https://optimized-einsum.readthedocs.io/en/stable/ opt einsums are faster and our results on Opacus benchmarking also corroborate it. Differential Revision: D37128344 fbshipit-source-id: 961fb1e487542a4226ea09367911fdf88252b934
sayanghosh
force-pushed
the
export-D37128344
branch
from
dcc4acf to
e55464d
Compare
|
Attached a notebook of the benchmarking and runtime comparison analysis here : https://colab.research.google.com/drive/1WAjSYzyKg7UisShNNwD4FMv6tIMGCzuL#scrollTo=OY0TWxFuXRwt |
karthikprasad
approved these changes
Jun 23, 2022
requirements.txt
Outdated
| @@ -1,3 +1,4 @@ | |||
| numpy>=1.15 | |||
| torch>=1.8 | |||
| scipy>=1.2 | |||
| opt-einsum==3.3.0 | |||
There was a problem hiding this comment.
Suggested change
| opt-einsum==3.3.0 | |
| opt-einsum>=3.3.0 |
sayanghosh
added a commit
to sayanghosh/opacus
that referenced
this pull request
Jun 24, 2022
Summary: Pull Request resolved: pytorch#440 We are using optimized einsums in place of Pytorch einsums. As per https://optimized-einsum.readthedocs.io/en/stable/ opt einsums are faster and our results on Opacus benchmarking also corroborate it. Differential Revision: D37128344 fbshipit-source-id: 37f73220311c7a7e161479868ed0a5da3c3f54f2
sayanghosh
force-pushed
the
export-D37128344
branch
from
e55464d to
3c6b7e7
Compare
|
This pull request was exported from Phabricator. Differential Revision: D37128344 |
sayanghosh
added a commit
to sayanghosh/opacus
that referenced
this pull request
Jun 24, 2022
Summary: Pull Request resolved: pytorch#440 We are using optimized einsums in place of Pytorch einsums. As per https://optimized-einsum.readthedocs.io/en/stable/ opt einsums are faster and our results on Opacus benchmarking also corroborate it. Differential Revision: D37128344 fbshipit-source-id: b03230542df2efa9036e52e8c869c91a79a1599e
|
This pull request was exported from Phabricator. Differential Revision: D37128344 |
sayanghosh
force-pushed
the
export-D37128344
branch
from
3c6b7e7 to
c317bb0
Compare
sayanghosh
added a commit
to sayanghosh/opacus
that referenced
this pull request
Jun 24, 2022
Summary: Pull Request resolved: pytorch#440 We are using optimized einsums in place of Pytorch einsums. As per https://optimized-einsum.readthedocs.io/en/stable/ opt einsums are faster and our results on Opacus benchmarking also corroborate it. Differential Revision: D37128344 fbshipit-source-id: e262b4a1199ea64b92c3bea60ddeb4c0363bce1d
|
This pull request was exported from Phabricator. Differential Revision: D37128344 |
sayanghosh
force-pushed
the
export-D37128344
branch
from
c317bb0 to
c8c8b0c
Compare
sayanghosh
added a commit
to sayanghosh/opacus
that referenced
this pull request
Jun 24, 2022
Summary: Pull Request resolved: pytorch#440 We are using optimized einsums in place of Pytorch einsums. As per https://optimized-einsum.readthedocs.io/en/stable/ opt einsums are faster and our results on Opacus benchmarking also corroborate it. Differential Revision: D37128344 fbshipit-source-id: 9afb0570011a48fce73ac41ee0497759e83f545b
sayanghosh
force-pushed
the
export-D37128344
branch
from
c8c8b0c to
a664e86
Compare
|
This pull request was exported from Phabricator. Differential Revision: D37128344 |
sayanghosh
added a commit
to sayanghosh/opacus
that referenced
this pull request
Jun 24, 2022
Summary: Pull Request resolved: pytorch#440 We are using optimized einsums in place of Pytorch einsums. As per https://optimized-einsum.readthedocs.io/en/stable/ opt einsums are faster and our results on Opacus benchmarking also corroborate it. Differential Revision: D37128344 fbshipit-source-id: 9f01084e2af026019dfa955c8b5f38de22a0845d
sayanghosh
force-pushed
the
export-D37128344
branch
from
a664e86 to
afe7bbb
Compare
This comment was marked as duplicate.
This comment was marked as duplicate.
sayanghosh
added a commit
to sayanghosh/opacus
that referenced
this pull request
Jun 24, 2022
Summary: Pull Request resolved: pytorch#440 We are using optimized einsums in place of Pytorch einsums. As per https://optimized-einsum.readthedocs.io/en/stable/ opt einsums are faster and our results on Opacus benchmarking also corroborate it. Differential Revision: D37128344 fbshipit-source-id: 891c1cc3e1348a4965a068d6fd1375eb584805b9
sayanghosh
force-pushed
the
export-D37128344
branch
from
afe7bbb to
f43033d
Compare
This comment was marked as duplicate.
This comment was marked as duplicate.
sayanghosh
added a commit
to sayanghosh/opacus
that referenced
this pull request
Jun 24, 2022
Summary: Pull Request resolved: pytorch#440 We are using optimized einsums in place of Pytorch einsums. As per https://optimized-einsum.readthedocs.io/en/stable/ opt einsums are faster and our results on Opacus benchmarking also corroborate it. Differential Revision: D37128344 fbshipit-source-id: 8753f406a325125a346530e57ef580d4a0abd1c9
sayanghosh
force-pushed
the
export-D37128344
branch
from
f43033d to
a4c62ea
Compare
This comment was marked as duplicate.
This comment was marked as duplicate.
Summary: Pull Request resolved: pytorch#440 We are using optimized einsums in place of Pytorch einsums. As per https://optimized-einsum.readthedocs.io/en/stable/ opt einsums are faster and our results on Opacus benchmarking also corroborate it. Reviewed By: karthikprasad Differential Revision: D37128344 fbshipit-source-id: 58bd594a349792f9112a356e3112f9bf5ded0c71
sayanghosh
force-pushed
the
export-D37128344
branch
from
a4c62ea to
106325f
Compare
|
This pull request was exported from Phabricator. Differential Revision: D37128344 |
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.
Differential Revision: D37128344