[quant][gpu][core] Implemented quantized add operator using cudnn [reland PR74463] #74463
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
Summary: This PR implements the quantized add operator using cudnn operations. A test case was added .... Test Plan: TBA [ghstack-poisoned]
CI Flow Status⚛️ CI FlowRuleset - Version:
|
Contributor
🔗 Helpful links
💊 CI failures summary and remediationsAs of commit d378f4c (more details on the Dr. CI page): 💚 💚 Looks good so far! There are no failures yet. 💚 💚 This comment was automatically generated by Dr. CI (expand for details).Please report bugs/suggestions to the (internal) Dr. CI Users group. |
Contributor
Author
|
@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator. |
… cudnn" Summary: This PR implements the quantized add operator using cudnn operations. A test case was added .... Test Plan: TBA Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
… cudnn" Summary: This PR implements the quantized add operator using cudnn operations. A test case was added .... Test Plan: TBA Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
… cudnn" Summary: This PR implements the quantized add operator using cudnn operations. A test case was added .... Test Plan: TBA Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
… cudnn" Summary: This PR implements the quantized add operator using cudnn operations. Added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
… cudnn" Summary: This PR implements the quantized add operator using cudnn operations. Added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
dzdang
added a commit
that referenced
this pull request
Mar 22, 2022
Summary: This PR implements the quantized add operator using cudnn operations. Added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` ghstack-source-id: 908412b Pull Request resolved: #74463
jerryzh168
reviewed
Mar 22, 2022
…udnn" Summary: This PR implements the quantized add operator using cudnn operations. Added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
dzdang
changed the title
dzdang
added a commit
that referenced
this pull request
Mar 22, 2022
Summary: This PR implements the quantized add operator using cudnn operations. Added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` ghstack-source-id: ef6fbf9 Pull Request resolved: #74463
|
We have recently simplified the CIFlow labels and
|
…udnn" Summary: This PR implements the quantized add operator using cudnn operations. Added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
…g cudnn" Summary: This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
…g cudnn" Summary: This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
dzdang
added a commit
that referenced
this pull request
Mar 25, 2022
Summary: This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` ghstack-source-id: 4de9f48 Pull Request resolved: #74463
Contributor
Author
|
@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator. |
1 similar comment
Contributor
Author
|
@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator. |
facebook-github-bot
pushed a commit
that referenced
this pull request
Mar 28, 2022
…4463) Summary: Pull Request resolved: #74463 This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` TBA Differential Revision: D35009111 D35009111 Reviewed By: jerryzh168 Pulled By: dzdang fbshipit-source-id: 13afa7f0192ffaf1f36334b1af827202c7dd0f74
jjsjann123
reviewed
Mar 28, 2022
| std::vector<int64_t> new_sizes(3, 1); | ||
| // cudnn expects leading dimensions to be the dummy dimensions | ||
| new_sizes.back() = qa.sizes().back(); | ||
| if (qa.ndim() == 2) { |
Collaborator
There was a problem hiding this comment.
This line should be qa.dim() instead, similar to previous line 67.
Contributor
|
This pull request has been reverted by 7bb0133. To re-land this change, please open another pull request, assignthe same reviewers, fix the CI failures that caused the revert and make sure that the failing CI runs on the PR by applying the proper ciflow label (e.g., ciflow/trunk). |
…g cudnn" Summary: This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
dzdang
added a commit
that referenced
this pull request
Mar 29, 2022
Summary: This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` ghstack-source-id: eb4c059 Pull Request resolved: #74463
…g cudnn" Summary: This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
dzdang
changed the title
…g cudnn [reland PR74463]" Summary: This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
dzdang
added a commit
that referenced
this pull request
Mar 29, 2022
…land PR74463] Summary: This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` ghstack-source-id: 4713a57 Pull Request resolved: #74463
…g cudnn [reland PR74463]" Summary: This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` a Differential Revision: [D35009111](https://our.internmc.facebook.com/intern/diff/D35009111) [ghstack-poisoned]
…g cudnn [reland PR74463]" Summary: This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` [ghstack-poisoned]
dzdang
added a commit
that referenced
this pull request
Mar 29, 2022
…land PR74463] Summary: This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` ghstack-source-id: 4713a57 Pull Request resolved: #74463
Contributor
Author
|
@dzdang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator. |
facebook-github-bot
pushed a commit
that referenced
this pull request
Mar 30, 2022
…land PR74463] (#74463) Summary: Pull Request resolved: #74463 This PR implements the quantized add operator using cudnn operations. Also added a corresponding test function in test_quantized_op.py. Ideally, we should merge this function with the cpu variant, but for now, we will keep it separate until cudnn v8 is in the default build. Other factors also complicate the merge as cudnn quantized add is currently only supported for int8 symmetrically quantized tensors. Test Plan: In pytorch main dir, execute ``` python test/test_quantization.py TestQuantizedOps.test_qadd_relu_cudnn ``` Reviewed By: ngimel Differential Revision: D35218224 Pulled By: dzdang fbshipit-source-id: a2e57e0b46cff655f2fb77000ea4db3a558a0851
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.
Stack from ghstack (oldest at bottom):
Summary:
This PR implements the quantized add operator using cudnn operations.
Also added a corresponding test function in test_quantized_op.py. Ideally,
we should merge this function with the cpu variant, but for now, we will
keep it separate until cudnn v8 is in the default build. Other factors also
complicate the merge as cudnn quantized add is currently only supported for
int8 symmetrically quantized tensors.
Test Plan:
In pytorch main dir, execute
Differential Revision: D35218224