Add optimized TBE training forward #1804
Closed
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 canceled.
|
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 5, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: e5b6ad27fc6a914a4def4d988946555daf0072a0
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 5, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: b961811488a25904a3f34660c553067b1ab93c95
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 5, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: 38944e9654ea06c3bec20b8c72d39a576f325367
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 5, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: 86328c01d296182f0b9180b05b32cdfc4c5ea750
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 6, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: cf6b6460194b7c6387dcfe82f2b7c6f4e8e06627
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 6, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: 232517acf727e0373876d1b31b45bbf059e7a2c0
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 7, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: c5281c474671ebe8efb0064db4a8fd2a97c3d917
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 7, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: c5130ffdbee8bb6d8e8112386635609738ea8a84
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 7, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: f1823db21a155794e6c2e7602ef7b37a7bd0ad59
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 7, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: df896e3aeb129d32a09cea6ac9873db43a96cbb3
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 8, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: e0530c653023eb57f408cf9b8d3d89486cc503f1
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 8, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: e441f0b4d4c76bf69aa5587e2e58766a13b59f4c
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
sryap
added a commit
to sryap/FBGEMM
that referenced
this pull request
Jun 12, 2023
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: f56b60a071f8c44b046de8a0c15d5d0f46dd1570
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
Summary: Pull Request resolved: pytorch#1804 This diff adds the frontend changes and tests for TBE v2 (D43634651) The `FBGEMM_EXPERIMENTAL_TBE` environment variable flag is added for enabling/disabling the new implementation at runtime. If `FBGEMM_EXPERIMENTAL_TBE` is not set, TBE will use the orignal implementation. If `FBGEMM_EXPERIMENTAL_TBE=1`, TBE will use the new implementation. If the TBE usecases are not supported in the new implementation, TBE will fall back to the original implementation. By default, `FBGEMM_EXPERIMENTAL_TBE` is not set. This can also be enabled by passing `use_experimental_tbe=True` when instantiating the TBE operator. ``` emb_op = SplitTableBatchedEmbeddingBagsCodegen( embedding_specs=..., ..., use_experimental_tbe=True, ) ``` Reviewed By: jianyuh Differential Revision: D44479772 fbshipit-source-id: 295fe8e28ba241cb640db15706dc15866f3effc8
|
This pull request was exported from Phabricator. Differential Revision: D44479772 |
|
This pull request has been merged in 38efe0d. |
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:
This diff adds the frontend changes and tests for TBE v2 (D43634651)
The
FBGEMM_EXPERIMENTAL_TBEenvironment variable flag is added forenabling/disabling the new implementation at runtime. If
FBGEMM_EXPERIMENTAL_TBEis not set, TBE will use the orignalimplementation. If
FBGEMM_EXPERIMENTAL_TBE=1, TBE will use the newimplementation. If the TBE usecases are not supported in the new
implementation, TBE will fall back to the original implementation. By
default,
FBGEMM_EXPERIMENTAL_TBEis not set.This can also be enabled by passing
use_experimental_tbe=Truewheninstantiating the TBE operator.
Differential Revision: D44479772