[Intel MKL] Supporting exponential_avg_factor attr for _MklFusedBatch… #37176
There are no files selected for viewing
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 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 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
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,259 @@ | ||
| /* Copyright 2020 The TensorFlow Authors. All Rights Reserved. | ||
|
|
||
| Licensed under the Apache License, Version 2.0 (the "License"); | ||
| you may not use this file except in compliance with the License. | ||
| You may obtain a copy of the License at | ||
|
|
||
| http://www.apache.org/licenses/LICENSE-2.0 | ||
|
|
||
| Unless required by applicable law or agreed to in writing, software | ||
| distributed under the License is distributed on an "AS IS" BASIS, | ||
| WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
| See the License for the specific language governing permissions and | ||
| limitations under the License. | ||
| ==============================================================================*/ | ||
| #ifdef INTEL_MKL | ||
|
There was a problem hiding this comment. Nit: Please add a newline before this line. |
||
| #include "tensorflow/cc/ops/const_op.h" | ||
| #include "tensorflow/cc/ops/image_ops.h" | ||
| #include "tensorflow/cc/ops/nn_ops.h" | ||
| #include "tensorflow/cc/ops/standard_ops.h" | ||
| #include "tensorflow/core/common_runtime/kernel_benchmark_testlib.h" | ||
| #include "tensorflow/core/framework/fake_input.h" | ||
| #include "tensorflow/core/framework/node_def_builder.h" | ||
| #include "tensorflow/core/framework/tensor.h" | ||
| #include "tensorflow/core/framework/types.pb.h" | ||
| #include "tensorflow/core/kernels/conv_ops_gpu.h" | ||
| #include "tensorflow/core/kernels/ops_testutil.h" | ||
| #include "tensorflow/core/kernels/ops_util.h" | ||
| #include "tensorflow/core/platform/test.h" | ||
| #include "tensorflow/core/platform/test_benchmark.h" | ||
| #include "tensorflow/core/platform/types.h" | ||
| #include "tensorflow/core/public/session.h" | ||
|
|
||
| namespace tensorflow { | ||
|
|
||
| // Helper class for converting MKL tensors to TF tensors and comparing to | ||
| // expected values | ||
|
|
||
| static const uint8 dummy_tensor[] = {0, 0, 0, 0, 0, 0, 0, 0}; | ||
| static const TensorShape dummy_shape({8}); | ||
|
|
||
| using GraphRunner = std::function<void( | ||
| const Tensor& input, const Tensor& scale, const Tensor& offset, | ||
| const Tensor& mean, const Tensor& variance, | ||
| const float exponential_avg_factor, const bool is_training, Tensor* output, | ||
| Tensor* batch_mean, Tensor* batch_var)>; | ||
|
|
||
| template <typename T> | ||
| class CommonTestUtilities : public OpsTestBase { | ||
| public: | ||
| void PerformConversion(DataType dtype, const Tensor& tensor, | ||
| const Tensor& mkl_meta_tensor, Tensor* output) { | ||
| // Create an MKL to TF conversion node and execute it | ||
| TF_EXPECT_OK(NodeDefBuilder("mkl_to_tf_op", "_MklToTf") | ||
| .Input(FakeInput(dtype)) // Input | ||
| .Input(FakeInput(DT_UINT8)) // Mkl second tensor | ||
| .Attr("T", dtype) | ||
| .Attr("_kernel", "MklLayoutDependentOp") | ||
| .Finalize(node_def())); | ||
| TF_EXPECT_OK(InitOp()); | ||
| AddInputFromArray<T>(tensor.shape(), tensor.flat<T>()); | ||
| AddInputFromArray<uint8>(mkl_meta_tensor.shape(), | ||
| mkl_meta_tensor.flat<uint8>()); | ||
| TF_ASSERT_OK(RunOpKernel()); | ||
|
|
||
| *output = *GetOutput(0); | ||
| } | ||
|
|
||
| void TestBody() {} | ||
|
|
||
| static void VerifyTensorsClose(const float exponential_avg_factor, | ||
| const bool is_training, const GraphRunner& run, | ||
| const GraphRunner& run_mkl) { | ||
| int batch = 1; | ||
| int height = 10; | ||
| int width = 10; | ||
| int depth = 3; | ||
| DataType dtype = DataTypeToEnum<T>::v(); | ||
|
|
||
| Tensor input(dtype, {batch, height, width, depth}); | ||
| input.flat<T>() = input.flat<T>().template setRandom<random_gen_>(); | ||
| Tensor scale(dtype, {depth}); | ||
| scale.flat<T>() = scale.flat<T>().template setRandom<random_gen_>(); | ||
| Tensor offset(dtype, {depth}); | ||
| offset.flat<T>() = offset.flat<T>().template setRandom<random_gen_>(); | ||
|
|
||
| if (is_training && (exponential_avg_factor == 1.0)) { | ||
| depth = 0; | ||
| } | ||
| Tensor mean(dtype, {depth}); | ||
| mean.flat<T>() = mean.flat<T>().template setRandom<random_gen_>(); | ||
| Tensor variance(dtype, {depth}); | ||
| variance.flat<T>() = | ||
| variance.flat<T>().template setRandom<random_gen_>().abs(); | ||
|
|
||
| Tensor output; | ||
| Tensor batch_mean; | ||
| Tensor batch_var; | ||
| Tensor mkl_output; | ||
| Tensor mkl_batch_mean; | ||
| Tensor mkl_batch_var; | ||
|
|
||
| run(input, scale, offset, mean, variance, exponential_avg_factor, | ||
| is_training, &output, &batch_mean, &batch_var); | ||
| run_mkl(input, scale, offset, mean, variance, exponential_avg_factor, | ||
| is_training, &mkl_output, &mkl_batch_mean, &mkl_batch_var); | ||
|
|
||
| ASSERT_EQ(output.dtype(), mkl_output.dtype()); | ||
| ASSERT_EQ(output.shape(), mkl_output.shape()); | ||
| ASSERT_EQ(batch_mean.dtype(), mkl_batch_mean.dtype()); | ||
| ASSERT_EQ(batch_mean.shape(), mkl_batch_mean.shape()); | ||
| ASSERT_EQ(batch_var.dtype(), mkl_batch_var.dtype()); | ||
| ASSERT_EQ(batch_var.shape(), mkl_batch_var.shape()); | ||
|
|
||
| test::ExpectClose(output, mkl_output, 1e-5); | ||
| test::ExpectClose(batch_mean, mkl_batch_mean, 1e-5); | ||
| test::ExpectClose(batch_var, mkl_batch_var, 1e-5); | ||
| } | ||
|
|
||
| private: | ||
| using random_gen_ = Eigen::internal::NormalRandomGenerator<T>; | ||
| }; | ||
|
|
||
| template <typename T> | ||
| class FusedBatchNormOpTest : public OpsTestBase { | ||
| protected: | ||
| void VerifyFusedBatchNorm(const float exponential_avg_factor, | ||
| const bool is_training) { | ||
| const GraphRunner run = [this](const Tensor& input, const Tensor& scale, | ||
| const Tensor& offset, const Tensor& mean, | ||
| const Tensor& variance, | ||
| const float exponential_avg_factor, | ||
| const bool is_training, Tensor* output, | ||
| Tensor* batch_mean, Tensor* batch_var) { | ||
| auto root = tensorflow::Scope::NewRootScope(); | ||
| auto input_op = | ||
| ops::Const(root.WithOpName("input"), Input::Initializer(input)); | ||
| auto scale_op = | ||
| ops::Const(root.WithOpName("scale"), Input::Initializer(scale)); | ||
| auto offset_op = | ||
| ops::Const(root.WithOpName("offset"), Input::Initializer(offset)); | ||
| auto mean_op = | ||
| ops::Const(root.WithOpName("mean"), Input::Initializer(mean)); | ||
| auto var_op = | ||
| ops::Const(root.WithOpName("variance"), Input::Initializer(variance)); | ||
|
|
||
| ops::FusedBatchNorm::Attrs attr; | ||
| attr = attr.IsTraining(is_training); | ||
| attr = attr.ExponentialAvgFactor(exponential_avg_factor); | ||
| attr = attr.Epsilon(0.001); | ||
| auto bn = ops::FusedBatchNorm(root.WithOpName("FusedBatchNorm"), input_op, | ||
| scale_op, offset_op, mean_op, var_op, attr); | ||
| auto y = ops::Identity(root.WithOpName("y"), bn.y); | ||
| auto y_batch_mean = | ||
| ops::Identity(root.WithOpName("y_batch_mean"), bn.batch_mean); | ||
| auto y_batch_var = | ||
| ops::Identity(root.WithOpName("y_batch_var"), bn.batch_variance); | ||
|
|
||
| tensorflow::GraphDef graph; | ||
| TF_ASSERT_OK(root.ToGraphDef(&graph)); | ||
|
|
||
| std::unique_ptr<tensorflow::Session> session( | ||
| tensorflow::NewSession(tensorflow::SessionOptions())); | ||
| TF_ASSERT_OK(session->Create(graph)); | ||
|
|
||
| std::vector<Tensor> output_tensors; | ||
| TF_ASSERT_OK(session->Run({}, {"y", "y_batch_mean", "y_batch_var"}, {}, | ||
| &output_tensors)); | ||
|
|
||
| *output = output_tensors[0]; | ||
| *batch_mean = output_tensors[1]; | ||
| *batch_var = output_tensors[2]; | ||
| }; | ||
|
|
||
| const GraphRunner run_mkl = [this](const Tensor& input, const Tensor& scale, | ||
| const Tensor& offset, const Tensor& mean, | ||
| const Tensor& variance, | ||
| const float exponential_avg_factor, | ||
| const bool is_training, Tensor* output, | ||
| Tensor* batch_mean, Tensor* batch_var) { | ||
| DataType dtype = DataTypeToEnum<T>::v(); | ||
| TF_EXPECT_OK(NodeDefBuilder("MklFusedBatchNorm", "_MklFusedBatchNorm") | ||
| .Input(FakeInput(dtype)) | ||
| .Input(FakeInput(DT_FLOAT)) | ||
| .Input(FakeInput(DT_FLOAT)) | ||
| .Input(FakeInput(DT_FLOAT)) | ||
| .Input(FakeInput(DT_FLOAT)) | ||
| .Input(FakeInput(DT_UINT8)) | ||
| .Input(FakeInput(DT_UINT8)) | ||
| .Input(FakeInput(DT_UINT8)) | ||
| .Input(FakeInput(DT_UINT8)) | ||
| .Input(FakeInput(DT_UINT8)) | ||
| .Attr("exponential_avg_factor", exponential_avg_factor) | ||
| .Attr("epsilon", 0.001) | ||
| .Attr("is_training", is_training) | ||
| .Attr("_kernel", "MklLayoutDependentOp") | ||
| .Finalize(node_def())); | ||
| TF_EXPECT_OK(InitOp()); | ||
|
|
||
| AddInputFromArray<float>(input.shape(), input.flat<T>()); | ||
| AddInputFromArray<float>(scale.shape(), scale.flat<T>()); | ||
| AddInputFromArray<float>(offset.shape(), offset.flat<T>()); | ||
| AddInputFromArray<float>(mean.shape(), mean.flat<T>()); | ||
| AddInputFromArray<float>(variance.shape(), variance.flat<T>()); | ||
| for (int i = 0; i < 5; ++i) | ||
| AddInputFromArray<uint8>(dummy_shape, dummy_tensor); | ||
| TF_ASSERT_OK(RunOpKernel()); | ||
|
|
||
| CommonTestUtilities<T> test_util; | ||
| test_util.PerformConversion(dtype, *GetOutput(0), *GetOutput(5), output); | ||
|
|
||
| CommonTestUtilities<T> test_util_mean; | ||
| test_util_mean.PerformConversion(dtype, *GetOutput(1), *GetOutput(6), | ||
| batch_mean); | ||
|
|
||
| CommonTestUtilities<T> test_util_var; | ||
| test_util_var.PerformConversion(dtype, *GetOutput(2), *GetOutput(7), | ||
| batch_var); | ||
| }; | ||
|
|
||
| CommonTestUtilities<T>::VerifyTensorsClose(exponential_avg_factor, | ||
| is_training, run, run_mkl); | ||
| } | ||
| }; | ||
|
|
||
| TYPED_TEST_CASE_P(FusedBatchNormOpTest); | ||
|
|
||
| TYPED_TEST_P(FusedBatchNormOpTest, Training) { | ||
| const float exponential_avg_factor = 1.0; | ||
| const bool is_training = true; | ||
| this->VerifyFusedBatchNorm(exponential_avg_factor, is_training); | ||
| } | ||
|
|
||
| TYPED_TEST_P(FusedBatchNormOpTest, TrainingRunningMean) { | ||
| const float exponential_avg_factor = 0.5; | ||
| const bool is_training = true; | ||
| this->VerifyFusedBatchNorm(exponential_avg_factor, is_training); | ||
| } | ||
|
|
||
| TYPED_TEST_P(FusedBatchNormOpTest, Inference) { | ||
| const float exponential_avg_factor = 1.0; | ||
| const bool is_training = false; | ||
| this->VerifyFusedBatchNorm(exponential_avg_factor, is_training); | ||
| } | ||
|
|
||
| TYPED_TEST_P(FusedBatchNormOpTest, InferenceIgnoreAvgFactor) { | ||
| const float exponential_avg_factor = 0.5; | ||
| const bool is_training = false; | ||
| this->VerifyFusedBatchNorm(exponential_avg_factor, is_training); | ||
| } | ||
|
|
||
| REGISTER_TYPED_TEST_CASE_P(FusedBatchNormOpTest, Training, TrainingRunningMean, | ||
|
There was a problem hiding this comment.
|
||
| Inference, InferenceIgnoreAvgFactor); | ||
|
|
||
| using FusedBatchNormDataTypes = ::testing::Types<float>; | ||
| INSTANTIATE_TYPED_TEST_CASE_P(Test, FusedBatchNormOpTest, | ||
|
There was a problem hiding this comment.
|
||
| FusedBatchNormDataTypes); | ||
|
|
||
| } // namespace tensorflow | ||
| #endif // INTEL_MKL | ||
|
There was a problem hiding this comment. Nit: Please add a newline before this line. |
||
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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can this be simplified to
?
If it can, but you'd rather use the original code to save testing time, please add a TODO.