Support int32_t indices/offsets for caching handling logics #811
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
|
This pull request was exported from Phabricator. Differential Revision: D33045589 |
jianyuh
added a commit
to jianyuh/FBGEMM
that referenced
this pull request
Dec 13, 2021
) Summary: Pull Request resolved: pytorch#811 In training, we assume the indices / offsets are int64_t for embedding (TBE), but in inference, we assume the indices / offsets are int32_t. This Diff enables both int32_t and int64_t supports for the caching logics so that we can reuse the same functions for both training and inference, while reducing the extra overhead to convert the indices/offsets from int to long or vice versa. Differential Revision: D33045589 fbshipit-source-id: 9f2f31110a0a070f7dadfe643b8585789303f145
jianyuh
force-pushed
the
export-D33045589
branch
from
4eb066c
to
c98d353
Compare
|
This pull request was exported from Phabricator. Differential Revision: D33045589 |
jianyuh
added a commit
to jianyuh/FBGEMM
that referenced
this pull request
Dec 13, 2021
) Summary: Pull Request resolved: pytorch#811 In training, we assume the indices / offsets are int64_t for embedding (TBE), but in inference, we assume the indices / offsets are int32_t. This Diff enables both int32_t and int64_t supports for the caching logics so that we can reuse the same functions for both training and inference, while reducing the extra overhead to convert the indices/offsets from int to long or vice versa. Differential Revision: D33045589 fbshipit-source-id: 65bb1de7503e5c076f54cb964f16aaf75f8c0047
jianyuh
force-pushed
the
export-D33045589
branch
from
c98d353
to
7a294a5
Compare
|
This pull request was exported from Phabricator. Differential Revision: D33045589 |
jianyuh
added a commit
to jianyuh/FBGEMM
that referenced
this pull request
Dec 28, 2021
) Summary: Pull Request resolved: pytorch#811 In training, we assume the indices / offsets are int64_t for embedding (TBE), but in inference, we assume the indices / offsets are int32_t. This Diff enables both int32_t and int64_t supports for the caching logics so that we can reuse the same functions for both training and inference, while reducing the extra overhead to convert the indices/offsets from int to long or vice versa. Differential Revision: D33045589 fbshipit-source-id: d2ccf90fe4d0a5ee40627bc9ec591c683fbfc993
jianyuh
force-pushed
the
export-D33045589
branch
from
7a294a5
to
d19ac06
Compare
|
This pull request was exported from Phabricator. Differential Revision: D33045589 |
jianyuh
added a commit
to jianyuh/FBGEMM
that referenced
this pull request
Dec 28, 2021
) Summary: Pull Request resolved: pytorch#811 In training, we assume the indices / offsets are int64_t for embedding (TBE), but in inference, we assume the indices / offsets are int32_t. This Diff enables both int32_t and int64_t supports for the caching logics so that we can reuse the same functions for both training and inference, while reducing the extra overhead to convert the indices/offsets from int to long or vice versa. Differential Revision: D33045589 fbshipit-source-id: 42ebcd899bb5dc6735eaf67cad48ac3b168d60ca
jianyuh
force-pushed
the
export-D33045589
branch
from
d19ac06
to
ec55e96
Compare
|
This pull request was exported from Phabricator. Differential Revision: D33045589 |
1 similar comment
|
This pull request was exported from Phabricator. Differential Revision: D33045589 |
jianyuh
added a commit
to jianyuh/FBGEMM
that referenced
this pull request
Dec 28, 2021
) Summary: Pull Request resolved: pytorch#811 In training, we assume the indices / offsets are int64_t for embedding (TBE), but in inference, we assume the indices / offsets are int32_t. This Diff enables both int32_t and int64_t supports for the caching logics so that we can reuse the same functions for both training and inference, while reducing the extra overhead to convert the indices/offsets from int to long or vice versa. Reviewed By: jspark1105 Differential Revision: D33045589 fbshipit-source-id: 4cdc7cec15e07c51af999276bf5366199eb216b5
jianyuh
force-pushed
the
export-D33045589
branch
from
ec55e96
to
ae22b20
Compare
jianyuh
added a commit
to jianyuh/FBGEMM
that referenced
this pull request
Dec 28, 2021
) Summary: Pull Request resolved: pytorch#811 In training, we assume the indices / offsets are int64_t for embedding (TBE), but in inference, we assume the indices / offsets are int32_t. This Diff enables both int32_t and int64_t supports for the caching logics so that we can reuse the same functions for both training and inference, while reducing the extra overhead to convert the indices/offsets from int to long or vice versa. Reviewed By: jspark1105 Differential Revision: D33045589 fbshipit-source-id: 89047f7cc09aee2f4ff18aa5ed3fbd0a86c16dd2
jianyuh
force-pushed
the
export-D33045589
branch
from
ae22b20
to
90e777e
Compare
|
This pull request was exported from Phabricator. Differential Revision: D33045589 |
jianyuh
added a commit
to jianyuh/FBGEMM
that referenced
this pull request
Dec 29, 2021
) Summary: Pull Request resolved: pytorch#811 In training, we assume the indices / offsets are int64_t for embedding (TBE), but in inference, we assume the indices / offsets are int32_t. This Diff enables both int32_t and int64_t supports for the caching logics so that we can reuse the same functions for both training and inference, while reducing the extra overhead to convert the indices/offsets from int to long or vice versa. Reviewed By: jspark1105 Differential Revision: D33045589 fbshipit-source-id: 5f8566797413989cf19d354f583f6a6fc87385bc
|
This pull request was exported from Phabricator. Differential Revision: D33045589 |
jianyuh
force-pushed
the
export-D33045589
branch
from
90e777e
to
8d43129
Compare
jianyuh
added a commit
to jianyuh/FBGEMM
that referenced
this pull request
Dec 29, 2021
) Summary: Pull Request resolved: pytorch#811 In training, we assume the indices / offsets are int64_t for embedding (TBE), but in inference, we assume the indices / offsets are int32_t. This Diff enables both int32_t and int64_t supports for the caching logics so that we can reuse the same functions for both training and inference, while reducing the extra overhead to convert the indices/offsets from int to long or vice versa. Reviewed By: jspark1105 Differential Revision: D33045589 fbshipit-source-id: 02fa0be6c6917d82258204f0de8d15266af29c77
jianyuh
force-pushed
the
export-D33045589
branch
from
8d43129
to
4a28321
Compare
|
This pull request was exported from Phabricator. Differential Revision: D33045589 |
1 similar comment
|
This pull request was exported from Phabricator. Differential Revision: D33045589 |
jianyuh
added a commit
to jianyuh/FBGEMM
that referenced
this pull request
Dec 30, 2021
) Summary: Pull Request resolved: pytorch#811 In training, we assume the indices / offsets are int64_t for embedding (TBE), but in inference, we assume the indices / offsets are int32_t. This Diff enables both int32_t and int64_t supports for the caching logics so that we can reuse the same functions for both training and inference, while reducing the extra overhead to convert the indices/offsets from int to long or vice versa. Reviewed By: jspark1105 Differential Revision: D33045589 fbshipit-source-id: 14a32f4c7d6ecb3673c7db6216d1771f995e5aa8
jianyuh
force-pushed
the
export-D33045589
branch
from
4a28321
to
b01c2b1
Compare
) Summary: Pull Request resolved: pytorch#811 In training, we assume the indices / offsets are int64_t for embedding (TBE), but in inference, we assume the indices / offsets are int32_t. This Diff enables both int32_t and int64_t supports for the caching logics so that we can reuse the same functions for both training and inference, while reducing the extra overhead to convert the indices/offsets from int to long or vice versa. Reviewed By: jspark1105 Differential Revision: D33045589 fbshipit-source-id: 6b96a8333c35ae1151149fb3c5c2ad11f7f6507d
jianyuh
force-pushed
the
export-D33045589
branch
from
b01c2b1
to
acdd4b2
Compare
|
This pull request was exported from Phabricator. Differential Revision: D33045589 |
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:
In training, we assume the indices / offsets are int64_t for embedding (TBE), but in inference, we assume the indices / offsets are int32_t.
This Diff enables both int32_t and int64_t supports for the caching logics so that we can reuse the same functions for both training and inference, while reducing the extra overhead to convert the indices/offsets from int to long or vice versa.
Differential Revision: D33045589