Creating ReadOnlyEmbeddingKVDB class and necessary functions #4225
Conversation
This file contains hidden or 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
✅ Deploy Preview for pytorch-fbgemm-docs ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
|
This pull request was exported from Phabricator. Differential Revision: D75489873 |
Raahul46
force-pushed
the
export-D75489873
branch
from
a94062d to
1ffac12
Compare
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 4, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Differential Revision: D75489873
Raahul46
force-pushed
the
export-D75489873
branch
from
1ffac12 to
71af63f
Compare
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 4, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Differential Revision: D75489873
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 4, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Differential Revision: D75489873
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 4, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Differential Revision: D75489873
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 4, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
|
This pull request was exported from Phabricator. Differential Revision: D75489873 |
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 4, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 4, 2025
…#4225) Summary: Pull Request resolved: pytorch#4225 X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
force-pushed
the
export-D75489873
branch
from
71af63f to
6c70898
Compare
|
This pull request was exported from Phabricator. Differential Revision: D75489873 |
Raahul46
force-pushed
the
export-D75489873
branch
from
6c70898 to
c358b86
Compare
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 4, 2025
…#4225) Summary: Pull Request resolved: pytorch#4225 X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
force-pushed
the
export-D75489873
branch
from
c358b86 to
6347c90
Compare
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
force-pushed
the
export-D75489873
branch
from
6347c90 to
669f83c
Compare
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
|
This pull request was exported from Phabricator. Differential Revision: D75489873 |
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: Pull Request resolved: pytorch#4225 X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
force-pushed
the
export-D75489873
branch
from
669f83c to
432090b
Compare
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
force-pushed
the
export-D75489873
branch
from
432090b to
34aad47
Compare
|
This pull request was exported from Phabricator. Differential Revision: D75489873 |
1 similar comment
|
This pull request was exported from Phabricator. Differential Revision: D75489873 |
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: Pull Request resolved: pytorch#4225 X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
force-pushed
the
export-D75489873
branch
from
34aad47 to
33da6c9
Compare
added 4 commits
June 5, 2025 10:29
Differential Revision: D75489819
Differential Revision: D75489827
Differential Revision: D75489841
Differential Revision: D75489849
Raahul46
force-pushed
the
export-D75489873
branch
from
33da6c9 to
cb2a2e7
Compare
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
force-pushed
the
export-D75489873
branch
from
cb2a2e7 to
9ab7366
Compare
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
|
This pull request was exported from Phabricator. Differential Revision: D75489873 |
Raahul46
pushed a commit
to Raahul46/FBGEMM-1
that referenced
this pull request
Jun 5, 2025
…#4225) Summary: Pull Request resolved: pytorch#4225 X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
Raahul46
force-pushed
the
export-D75489873
branch
from
9ab7366 to
2cc9cb7
Compare
…#4225) Summary: Pull Request resolved: pytorch#4225 X-link: facebookresearch/FBGEMM#1301 Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54 Context: We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold In this diff: The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently. To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class. We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object Reviewed By: duduyi2013 Differential Revision: D75489873
|
This pull request was exported from Phabricator. Differential Revision: D75489873 |
Raahul46
force-pushed
the
export-D75489873
branch
from
2cc9cb7 to
886e16f
Compare
|
This pull request has been merged in 2283ec6. |
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:
X-link: https://github.com/facebookresearch/FBGEMM/pull/1301
Design doc: https://docs.google.com/document/d/149LdAEHOLP7ei4hwVVkAFXGa4N9uLs1J7efxfBZp3dY/edit?tab=t.0#heading=h.49t3yfaqmt54
Context:
We are enabling the usage of rocksDB checkpoint feature in KVTensorWrapper. This allows us to create checkpoints of the embedding tables in SSD. Later, these checkpoints are used by the checkpointing component to create a checkpoint and upload it it to the manifold
In this diff:
The primary objective of adding the checkpointhandle is to allow multiple process read through the KVTensor. To enable this, we would require to create a read-only KVTensor object that can be read concurrently.
To support this, we introduce an ReadOnlyEmbedding KVDB class which is a read-only implementation of EmbeddingKVDB class.
We have added a new constructor to the KVTensorWrapper which takes in a serialized KVTensor meta data. When deserializing, we create a readOnlyEmbeddingKVDB for the KVTensorWrapper object
Differential Revision: D75489873