Merge branch 'mbencer/BuilderSplitV1' of…

Merge branch 'mbencer/BuilderSplitV1' of https://github.com/NervanaSystems/ngraph into mbencer/BuilderSplitV1

CMakeLists.txt

@@ -143,7 +143,7 @@ option(NGRAPH_PLAIDML_STATIC_LIB_ENABLE "Enable build PlaidML backend as a stati
 option(NGRAPH_DYNAMIC_COMPONENTS_ENABLE "Enable dynamic loading of components" TRUE)
 option(NGRAPH_NATIVE_ARCH_ENABLE "Enable build for native archtecture" TRUE)
 option(NGRAPH_EXPORT_TARGETS_ENABLE "Enable exporting nGraph cmake export targets" TRUE)
-option(NGRAPH_WARNINGS_AS_ERRORS "Make all nGraph compile-time warnings into errors" TRUE)
+option(NGRAPH_WARNINGS_AS_ERRORS "Make all nGraph compile-time warnings into errors" FALSE)
 if (NGRAPH_CPU_ENABLE)
     option(NGRAPH_TBB_ENABLE "Control usage of TBB for CPU backend" TRUE)
 endif()

CODEOWNERS

@@ -20,7 +20,7 @@
 /doc/                               @indie
 /doc/examples/mnist_mlp/dist_*      @wenzhe-nrv @indie
 /doc/*/*/frameworks/tensorflow_connect.rst      @shresthamalik @sayantan-nervana
-/doc/*/*/backends/plaidml-ng-api/   @flaub @brianretford @dgkutnic
+/doc/*/*/backends/plaidml-ng-api/   @diyessi
 /doc/*/*/inspection/                @diyessi
 /doc/examples/onnx/                 @arogowie-intel @tsocha
 /README.md                          @indie
@@ -55,7 +55,7 @@ project/doc-contributor-README.rst  @indie
 /src/ngraph/runtime/dynamic/        @diyessi
 /src/ngraph/runtime/gpu/            @csullivan @rkimballn1
 /src/ngraph/runtime/interpreter/    @rkimballn1
-/src/ngraph/runtime/plaidml/        @earhart @dgkutnic
+/src/ngraph/runtime/plaidml/        @diyessi
 /src/ngraph/runtime/reference/      @diyessi
 /src/ngraph/runtime/reference/allreduce.*pp      @wenzhe-nrv @diyessi
 /src/ngraph/type/                   @diyessi

src/contrib/mlir/core/pass/mlir_subgraph_extraction.cpp

@@ -20,6 +20,7 @@
 #include "mlir_subgraph_extraction.hpp"
 #include "ngraph/assertion.hpp"
 #include "ngraph/graph_util.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/node.hpp"
 #include "ngraph/ops.hpp"
 

src/ngraph/env_util.cpp

@@ -17,6 +17,7 @@
 #include <sstream>
 
 #include "ngraph/env_util.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/util.hpp"
 
 using namespace std;

src/ngraph/frontend/onnx_import/op/matmul_integer.cpp

@@ -16,6 +16,7 @@
 
 #include "matmul_integer.hpp"
 #include "ngraph/builder/matmul_factory.hpp"
+#include "ngraph/log.hpp"
 
 namespace ngraph
 {

src/ngraph/frontend/onnx_import/op/qlinear_matmul.cpp

@@ -16,6 +16,7 @@
 
 #include "qlinear_matmul.hpp"
 #include "ngraph/builder/matmul_factory.hpp"
+#include "ngraph/log.hpp"
 
 namespace ngraph
 {

src/ngraph/log.cpp

@@ -22,6 +22,7 @@
 #include <mutex>
 #include <thread>
 
+#include "ngraph/distributed.hpp"
 #include "ngraph/env_util.hpp"
 #include "ngraph/log.hpp"
 

src/ngraph/log.hpp

@@ -30,8 +30,6 @@
 #endif
 #include <vector>
 
-#include "ngraph/distributed.hpp"
-
 namespace ngraph
 {
     class ConstString

src/ngraph/op/allreduce.hpp

@@ -17,6 +17,7 @@
 #pragma once
 
 #include <memory>
+#include "ngraph/distributed.hpp"
 #include "ngraph/op/op.hpp"
 
 namespace ngraph

src/ngraph/op/avg_pool.cpp

@@ -15,6 +15,7 @@
 //*****************************************************************************
 
 #include "ngraph/op/avg_pool.hpp"
+#include "ngraph/attribute_visitor.hpp"
 #include "ngraph/graph_util.hpp"
 #include "ngraph/validation_util.hpp"
 

src/ngraph/op/batch_norm.cpp

@@ -16,6 +16,7 @@
 
 #include <sstream>
 
+#include "ngraph/attribute_visitor.hpp"
 #include "ngraph/op/batch_norm.hpp"
 #include "ngraph/op/get_output_element.hpp"
 #include "ngraph/validation_util.hpp"

src/ngraph/op/binary_convolution.cpp

@@ -15,6 +15,7 @@
 //*****************************************************************************
 
 #include "ngraph/op/binary_convolution.hpp"
+#include "ngraph/attribute_visitor.hpp"
 #include "ngraph/axis_vector.hpp"
 #include "ngraph/coordinate_diff.hpp"
 #include "ngraph/op/reshape.hpp"

src/ngraph/op/broadcast.cpp

@@ -15,6 +15,7 @@
 //*****************************************************************************
 
 #include "ngraph/op/broadcast.hpp"
+#include "ngraph/attribute_visitor.hpp"
 #include "ngraph/op/constant.hpp"
 #include "ngraph/op/sum.hpp"
 #include "ngraph/op/util/broadcasting.hpp"

src/ngraph/op/broadcast_distributed.cpp

@@ -15,6 +15,7 @@
 //*****************************************************************************
 
 #include "ngraph/op/broadcast_distributed.hpp"
+#include "ngraph/attribute_visitor.hpp"
 
 using namespace std;
 using namespace ngraph;

src/ngraph/op/concat.cpp

@@ -16,6 +16,7 @@
 
 #include <memory>
 
+#include "ngraph/attribute_visitor.hpp"
 #include "ngraph/op/concat.hpp"
 #include "ngraph/op/slice.hpp"
 

src/ngraph/op/fused/squeeze.cpp

@@ -32,20 +32,20 @@ op::Squeeze::Squeeze(const Output<Node>& data, const Output<Node>& axes)
     constructor_validate_and_infer_types();
 }
 
-NodeVector op::Squeeze::decompose_op() const
+void op::Squeeze::pre_validate_and_infer_types()
 {
     auto data = input_value(0);
     auto axes_node = input_value(1).get_node_shared_ptr();
 
-    // Currently only support Constant node for axes.
-    NODE_VALIDATION_CHECK(this,
-                          axes_node->is_constant(),
-                          "doesn't support 'axes' input of other type than a Constant.");
+    if (data.get_partial_shape().is_dynamic() || !axes_node->is_constant())
+    {
+        set_output_type(0, get_input_element_type(0), PartialShape::dynamic());
+        return;
+    }
 
     // Get value of axes from Constant
     auto axes_constant = as_type_ptr<op::Constant>(axes_node);
     auto axes = axes_constant->cast_vector<size_t>();
-
     auto data_shape = data.get_shape();
     std::vector<uint64_t> axes_to_squeeze(data_shape.size());
 
@@ -87,6 +87,18 @@ NodeVector op::Squeeze::decompose_op() const
         }
     }
 
+    set_output_type(0, get_input_element_type(0), output_data_shape);
+}
+
+NodeVector op::Squeeze::decompose_op() const
+{
+    NODE_VALIDATION_CHECK(
+        this,
+        (get_output_partial_shape(0).is_static()),
+        "output shape was not calculated during pre_validate_and_infer_types. Can not decompose.");
+    auto data = input_value(0);
+    auto data_shape = data.get_shape();
+    auto output_data_shape = get_output_shape(0);
     AxisVector input_order{get_default_order(data_shape.size())};
     return {make_shared<op::Reshape>(data, input_order, output_data_shape)};
 }

src/ngraph/op/fused/squeeze.hpp

@@ -38,6 +38,7 @@ namespace ngraph
                 Squeeze(const Output<Node>& data, const Output<Node>& axes);
 
                 virtual NodeVector decompose_op() const override;
+                virtual void pre_validate_and_infer_types() override;
 
                 virtual std::shared_ptr<Node>
                     copy_with_new_args(const NodeVector& new_args) const override;

src/ngraph/op/fused/unsqueeze.cpp

@@ -33,31 +33,35 @@ op::Unsqueeze::Unsqueeze(const Output<Node>& data, const Output<Node>& axes)
 }
 
 void op::Unsqueeze::pre_validate_and_infer_types()
-{
-    auto axes_node = input_value(1).get_node_shared_ptr();
-
-    // Currently only support Constant node for axes.
-    NODE_VALIDATION_CHECK(this,
-                          axes_node->is_constant(),
-                          "doesn't support 'axes' input of other type than a Constant.");
-}
-
-NodeVector op::Unsqueeze::decompose_op() const
 {
     auto data = input_value(0);
     auto axes_node = input_value(1).get_node_shared_ptr();
 
+    if (data.get_partial_shape().rank().is_dynamic() || !axes_node->is_constant())
+    {
+        set_output_type(0, get_input_element_type(0), PartialShape::dynamic());
+        return;
+    }
+
     // Get value of axes from Constant
     auto axes_constant = as_type_ptr<op::Constant>(axes_node);
     auto axes = axes_constant->cast_vector<size_t>();
 
-    auto data_shape = data.get_shape();
-
     NODE_VALIDATION_CHECK(this, !axes.empty(), "'axes' input is mandatory.");
     NODE_VALIDATION_CHECK(this,
                           axes.size() == set<int64_t>(begin(axes), end(axes)).size(),
                           "'axes' input has a duplicate axis.");
 
+    if (data.get_partial_shape().is_dynamic())
+    {
+        set_output_type(0,
+                        get_input_element_type(0),
+                        PartialShape::dynamic(data.get_partial_shape().rank() + axes.size()));
+        return;
+    }
+
+    auto data_shape = data.get_shape();
+
     sort(begin(axes), end(axes), less<int64_t>());
 
     AxisVector input_order{ngraph::get_default_order(data_shape.size())};
@@ -69,8 +73,20 @@ NodeVector op::Unsqueeze::decompose_op() const
 
         data_shape.insert(next(begin(data_shape), axis), 1);
     }
+    set_output_type(0, get_input_element_type(0), data_shape);
+}
 
-    return {make_shared<ngraph::op::Reshape>(data, input_order, data_shape)};
+NodeVector op::Unsqueeze::decompose_op() const
+{
+    NODE_VALIDATION_CHECK(
+        this,
+        (get_output_partial_shape(0).is_static()),
+        "output shape was not calculated during pre_validate_and_infer_types. Can not decompose.");
+    auto data = input_value(0);
+    auto data_shape = data.get_shape();
+    auto output_shape = get_output_shape(0);
+    AxisVector input_order{ngraph::get_default_order(data_shape.size())};
+    return {make_shared<ngraph::op::Reshape>(data, input_order, output_shape)};
 }
 
 shared_ptr<Node> op::Unsqueeze::copy_with_new_args(const NodeVector& new_args) const

src/ngraph/op/select.cpp

@@ -16,6 +16,7 @@
 
 #include <memory>
 
+#include "ngraph/attribute_visitor.hpp"
 #include "ngraph/log.hpp"
 #include "ngraph/op/convert.hpp"
 #include "ngraph/op/multiply.hpp"

src/ngraph/op/util/binary_elementwise_arithmetic.cpp

@@ -15,6 +15,7 @@
 //*****************************************************************************
 
 #include "ngraph/op/util/binary_elementwise_arithmetic.hpp"
+#include "ngraph/attribute_visitor.hpp"
 
 using namespace std;
 using namespace ngraph;

src/ngraph/op/util/binary_elementwise_comparison.cpp

@@ -15,6 +15,7 @@
 //*****************************************************************************
 
 #include "ngraph/op/util/binary_elementwise_comparison.hpp"
+#include "ngraph/attribute_visitor.hpp"
 
 using namespace std;
 using namespace ngraph;

src/ngraph/op/util/binary_elementwise_logical.cpp

@@ -15,6 +15,7 @@
 //*****************************************************************************
 
 #include "ngraph/op/util/binary_elementwise_logical.hpp"
+#include "ngraph/attribute_visitor.hpp"
 
 using namespace std;
 using namespace ngraph;

src/ngraph/op/util/index_reduction.cpp

@@ -16,6 +16,7 @@
 
 #include <memory>
 
+#include "ngraph/attribute_visitor.hpp"
 #include "ngraph/op/util/index_reduction.hpp"
 
 using namespace std;

src/ngraph/pass/batch_fusion.cpp

@@ -25,6 +25,7 @@
 #include "batch_fusion.hpp"
 
 #include "ngraph/graph_util.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/op/add.hpp"
 #include "ngraph/op/broadcast.hpp"
 #include "ngraph/op/concat.hpp"

src/ngraph/pass/constant_folding.hpp

@@ -16,6 +16,7 @@
 
 #pragma once
 
+#include "ngraph/log.hpp"
 #include "ngraph/pass/graph_rewrite.hpp"
 #include "ngraph/runtime/aligned_buffer.hpp"
 #include "ngraph/util.hpp"

src/ngraph/pass/pass_util.cpp

@@ -15,6 +15,7 @@
 //*****************************************************************************
 
 #include "ngraph/pass/pass_util.hpp"
+#include "ngraph/log.hpp"
 
 using namespace std;
 using namespace ngraph;

src/ngraph/pass/propagate_cacheability.cpp

@@ -17,6 +17,7 @@
 #include "ngraph/pass/propagate_cacheability.hpp"
 
 #include "ngraph/graph_util.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/op/constant.hpp"
 #include "ngraph/op/parameter.hpp"
 #include "ngraph/op/util/op_annotations.hpp"

src/ngraph/placement.hpp

@@ -22,8 +22,6 @@
 #include <unordered_set>
 #include <vector>
 
-#include "ngraph/log.hpp"
-
 namespace ngraph
 {
     enum class Placement

src/ngraph/runtime/cpu/builder/broadcast_distributed.cpp

@@ -15,6 +15,8 @@
 //*****************************************************************************
 
 #include "ngraph/op/broadcast_distributed.hpp"
+#include "ngraph/distributed.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/runtime/cpu/cpu_builder.hpp"
 
 using namespace std;

src/ngraph/runtime/cpu/builder/softmax.cpp

@@ -15,6 +15,7 @@
 //*****************************************************************************
 
 #include "ngraph/op/softmax.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/runtime/cpu/cpu_builder.hpp"
 #include "ngraph/runtime/cpu/kernel/softmax.hpp"
 #include "ngraph/runtime/cpu/mkldnn_invoke.hpp"

src/ngraph/runtime/cpu/cpu_backend.cpp

@@ -22,6 +22,7 @@
 
 #include "ngraph/component_manager.hpp"
 #include "ngraph/graph_util.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/runtime/backend_manager.hpp"
 #include "ngraph/runtime/cpu/cpu_backend.hpp"
 #include "ngraph/runtime/cpu/cpu_builder_registry.hpp"

src/ngraph/runtime/cpu/cpu_cse.cpp

@@ -15,6 +15,7 @@
 //*****************************************************************************
 
 #include "cpu_cse.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/op/reshape.hpp"
 #include "ngraph/runtime/cpu/cpu_layout_descriptor.hpp"
 #include "ngraph/runtime/cpu/cpu_op_annotations.hpp"

src/ngraph/runtime/cpu/cpu_visualize_tree.cpp

@@ -16,6 +16,7 @@
 
 #include <string>
 
+#include "ngraph/log.hpp"
 #include "ngraph/op/reshape.hpp"
 #include "ngraph/runtime/cpu/cpu_layout_descriptor.hpp"
 #include "ngraph/runtime/cpu/cpu_op_annotations.hpp"

src/ngraph/runtime/cpu/op/deconv.cpp

@@ -18,6 +18,7 @@
 
 #include "deconv.hpp"
 
+#include "ngraph/log.hpp"
 #include "ngraph/op/convolution.hpp"
 #include "ngraph/op/get_output_element.hpp"
 #include "ngraph/util.hpp"

src/ngraph/runtime/cpu/pass/cpu_assignment.cpp

@@ -24,6 +24,7 @@
 #include <mkldnn.hpp>
 
 #include "ngraph/descriptor/output.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/op/add.hpp"
 #include "ngraph/op/avg_pool.hpp"
 #include "ngraph/op/batch_norm.hpp"

src/ngraph/runtime/cpu/pass/cpu_mat_fusion.cpp

@@ -24,6 +24,7 @@
 
 #include "cpu_mat_fusion.hpp"
 #include "ngraph/graph_util.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/op/add.hpp"
 #include "ngraph/op/broadcast.hpp"
 #include "ngraph/op/concat.hpp"

src/ngraph/runtime/cpu/pass/cpu_memory_optimization.cpp

@@ -53,6 +53,7 @@
 
 #include "ngraph/descriptor/output.hpp"
 #include "ngraph/graph_util.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/op/concat.hpp"
 #include "ngraph/op/constant.hpp"
 #include "ngraph/op/slice.hpp"

src/ngraph/runtime/plaidml/plaidml_pass_explicit_logicals.cpp

@@ -17,6 +17,7 @@
 #include <typeindex>
 
 #include "ngraph/graph_util.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/op/and.hpp"
 #include "ngraph/op/equal.hpp"
 #include "ngraph/op/greater.hpp"

src/ngraph/serializer.cpp

@@ -23,6 +23,7 @@
 #include "ngraph/env_util.hpp"
 #include "ngraph/file_util.hpp"
 #include "ngraph/graph_util.hpp"
+#include "ngraph/log.hpp"
 #include "ngraph/ops.hpp"
 #include "ngraph/provenance.hpp"
 #include "ngraph/serializer.hpp"

test/bfloat16.cpp

@@ -19,6 +19,7 @@
 
 #include "gtest/gtest.h"
 
+#include "ngraph/log.hpp"
 #include "ngraph/runtime/aligned_buffer.hpp"
 #include "ngraph/type/bfloat16.hpp"
 #include "util/float_util.hpp"

test/type_prop/squeeze.cpp

@@ -37,3 +37,21 @@ TEST(type_prop, squeeze)
     ASSERT_EQ(squeeze_default_axes->get_element_type(), element::f32);
     ASSERT_EQ(squeeze_default_axes->get_shape(), (Shape{4, 4, 8}));
 }
+
+TEST(type_prop, squeeze_dynamic)
+{
+    auto param = make_shared<op::Parameter>(element::f32, PartialShape::dynamic(6));
+    auto axes_node =
+        make_shared<ngraph::op::Constant>(element::u64, Shape{2}, vector<int64_t>{0, 2});
+    auto squeeze = make_shared<op::Squeeze>(param, axes_node);
+
+    ASSERT_EQ(squeeze->get_element_type(), element::f32);
+    EXPECT_TRUE(squeeze->get_output_partial_shape(0).same_scheme(PartialShape::dynamic()));
+
+    axes_node = make_shared<ngraph::op::Constant>(element::u64, Shape{0}, vector<int64_t>{});
+    auto squeeze_default_axes = make_shared<op::Squeeze>(param, axes_node);
+
+    ASSERT_EQ(squeeze_default_axes->get_element_type(), element::f32);
+    EXPECT_TRUE(
+        squeeze_default_axes->get_output_partial_shape(0).same_scheme(PartialShape::dynamic()));
+}

test/type_prop/unsqueeze.cpp

@@ -26,8 +26,19 @@ TEST(type_prop, unsqueeze)
     auto param = make_shared<op::Parameter>(element::f32, Shape{4, 1, 4, 1, 8});
     auto axes_node =
         make_shared<ngraph::op::Constant>(element::u64, Shape{2}, vector<int64_t>{1, 2});
-    auto squeeze = make_shared<op::Unsqueeze>(param, axes_node);
+    auto unsqueeze = make_shared<op::Unsqueeze>(param, axes_node);
 
-    ASSERT_EQ(squeeze->get_element_type(), element::f32);
-    ASSERT_EQ(squeeze->get_shape(), (Shape{4, 1, 1, 1, 4, 1, 8}));
+    ASSERT_EQ(unsqueeze->get_element_type(), element::f32);
+    ASSERT_EQ(unsqueeze->get_shape(), (Shape{4, 1, 1, 1, 4, 1, 8}));
+}
+
+TEST(type_prop, unsqueeze_dynamic)
+{
+    auto param = make_shared<op::Parameter>(element::f32, PartialShape::dynamic(5));
+    auto axes_node =
+        make_shared<ngraph::op::Constant>(element::u64, Shape{2}, vector<int64_t>{1, 2});
+    auto unsqueeze = make_shared<op::Unsqueeze>(param, axes_node);
+
+    ASSERT_EQ(unsqueeze->get_element_type(), element::f32);
+    EXPECT_TRUE(unsqueeze->get_output_partial_shape(0).same_scheme(PartialShape::dynamic(7)));
 }