[quant] Add a quantized batch_norm operator #33080
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
Summary: Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
💊 CircleCI build failures summary and remediationsAs of commit 12507a3: Commit 12507a3 was recently pushed. Waiting for builds... This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions on the GitHub issue tracker. This comment has been revised 14 times. |
Summary: Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
raghuramank100
suggested changes
Feb 7, 2020
Contributor
raghuramank100
left a comment
raghuramank100
left a comment
There was a problem hiding this comment.
Looks good! A few requests:
jamesr66a
reviewed
Feb 7, 2020
aten/src/ATen/native/quantized/cpu/kernels/QuantizedOpKernels.cpp
Outdated
Show resolved
Hide resolved
aten/src/ATen/native/quantized/cpu/kernels/QuantizedOpKernels.cpp
Outdated
Show resolved
Hide resolved
aten/src/ATen/native/quantized/cpu/kernels/QuantizedOpKernels.cpp
Outdated
Show resolved
Hide resolved
aten/src/ATen/native/quantized/cpu/kernels/QuantizedOpKernels.cpp
Outdated
Show resolved
Hide resolved
Summary: Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
jamesr66a
reviewed
Feb 8, 2020
| constexpr int kVLen = 8; | ||
| const int outer_size = N * HxW; | ||
| using Vec = Vec256<quint8>; | ||
| for (int i = 0; i < outer_size; ++i) { |
Collaborator
There was a problem hiding this comment.
I think we might be able to simplify this further, since--correct me if i'm wrong--this is essentially just an elementwise access pattern where we broadcast alpha and beta to all [N, H, W]. It might be pretty simple to just use TensorIterator instead of handling the loop nest explicitly. I can elaborate a bit more later
Summary: Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
jamesr66a
approved these changes
Feb 12, 2020
aten/src/ATen/native/quantized/cpu/kernels/QuantizedOpKernels.cpp
Outdated
Show resolved
Hide resolved
Summary: Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D19861927](https://our.internmc.facebook.com/intern/diff/D19861927) [ghstack-poisoned]
Summary: Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D19861927](https://our.internmc.facebook.com/intern/diff/D19861927) [ghstack-poisoned]
Contributor
|
This pull request has been merged in d043560. |
ttumiel
pushed a commit
to ttumiel/pytorch
that referenced
this pull request
Mar 4, 2020
Summary: Pull Request resolved: pytorch#33080 Quantized batch norm for cases where batch norm cannot be fused with conv. AVX2 implementation is from Caffe2. Test Plan: python test/test_quantized.py TestQuantizedOps.test_batch_norm Imported from OSS Differential Revision: D19861927 fbshipit-source-id: bd8cd101fc063cb6358132ab7c651a160999293c
vkuzo
added a commit
that referenced
this pull request
Mar 24, 2020
Summary: Adds a quantized implementation of LayerNorm for server. Relevant PRs: * #20345 (floating point LN) * #33080 (quantized BN) A future PR will add the Python wrapper. Test Plan: numerics match the floating point implementation TODO: benchmarks Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
vkuzo
added a commit
that referenced
this pull request
Mar 24, 2020
Summary: Adds a quantized implementation of LayerNorm for server. Relevant PRs: * #20345 (floating point LN) * #33080 (quantized BN) A future PR will add the Python wrapper. Test Plan: numerics match the floating point implementation TODO: benchmarks Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
vkuzo
added a commit
that referenced
this pull request
Mar 24, 2020
Summary: Adds a quantized implementation of LayerNorm for server. Relevant PRs: * #20345 (floating point LN) * #33080 (quantized BN) A future PR will add the Python wrapper. Test Plan: numerics match the floating point implementation TODO: benchmarks Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
vkuzo
added a commit
that referenced
this pull request
Mar 24, 2020
Summary: Adds a quantized implementation of LayerNorm for server. Relevant PRs: * #20345 (floating point LN) * #33080 (quantized BN) A future PR will add the Python wrapper. Test Plan: numerics match the floating point implementation TODO: benchmarks Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 3c3721f Pull Request resolved: #35329
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:
Summary:
Quantized batch norm for cases where batch norm cannot be fused with conv.
AVX2 implementation is from Caffe2.
Test Plan:
python test/test_quantized.py TestQuantizedOps.test_batch_norm
Reviewers:
Subscribers:
Tasks:
Tags:
Differential Revision: D19861927