Merge pull request #164 from ROCmSoftwarePlatform/pad_op

asymmetric padding and autopad SAME_UPPER support

src/include/migraphx/operators.hpp

@@ -16,6 +16,13 @@ namespace migraphx {
 inline namespace MIGRAPHX_INLINE_NS {
 namespace op {
 
+enum padding_mode_t
+{
+    default_, // NOLINT
+    same,
+    valid
+};
+
 struct not_computable
 {
     argument compute(const shape&, const std::vector<argument>&) const
@@ -58,12 +65,7 @@ struct convolution
     std::array<std::size_t, 2> padding  = {{0, 0}};
     std::array<std::size_t, 2> stride   = {{1, 1}};
     std::array<std::size_t, 2> dilation = {{1, 1}};
-    enum padding_mode_t
-    {
-        default_, // NOLINT
-        same,
-        valid
-    };
+
     padding_mode_t padding_mode = default_;
     int group                   = 1;
 
@@ -138,12 +140,7 @@ struct im2col
     std::array<std::size_t, 2> padding  = {{0, 0}};
     std::array<std::size_t, 2> stride   = {{1, 1}};
     std::array<std::size_t, 2> dilation = {{1, 1}};
-    enum padding_mode_t
-    {
-        default_, // NOLINT
-        same,
-        valid
-    };
+
     padding_mode_t padding_mode = default_;
 
     template <class Self, class F>
@@ -189,12 +186,14 @@ struct pooling
     std::array<std::size_t, 2> padding = {{0, 0}};
     std::array<std::size_t, 2> stride  = {{1, 1}};
     std::array<std::size_t, 2> lengths = {{1, 1}};
+    padding_mode_t padding_mode        = default_;
 
     template <class Self, class F>
     static auto reflect(Self& self, F f)
     {
         return pack(f(self.mode, "mode"),
                     f(self.padding, "padding"),
+                    f(self.padding, "padding_mode"),
                     f(self.stride, "stride"),
                     f(self.lengths, "lengths"));
     }
@@ -211,7 +210,10 @@ struct pooling
         assert(lengths[0] <= (input.lens()[2] + 2 * padding[0]));
         assert(lengths[1] <= (input.lens()[3] + 2 * padding[1]));
 
-        return {t,
+        if(padding_mode == default_)
+        {
+            return {
+                t,
                 {
                     input.lens()[0],
                     input.lens()[1],
@@ -227,6 +229,39 @@ struct pooling
                             1)),
                 }};
         }
+        else if(padding_mode == same)
+        {
+            return {t,
+                    {input.lens()[0],
+                     input.lens()[1],
+                     static_cast<std::size_t>(
+                         std::ceil(static_cast<double>(input.lens()[2]) / stride[0])),
+                     static_cast<std::size_t>(
+                         std::ceil(static_cast<double>(input.lens()[3]) / stride[1]))}};
+        }
+        else if(padding_mode == valid)
+        {
+            return {t,
+                    {
+                        input.lens()[0],
+                        input.lens()[1],
+                        std::size_t(std::max<std::ptrdiff_t>(
+                            1,
+                            std::ptrdiff_t(std::floor((input.lens()[2] - lengths[0]) /
+                                                      static_cast<float>(stride[0]))) +
+                                1)),
+                        std::size_t(std::max<std::ptrdiff_t>(
+                            1,
+                            std::ptrdiff_t(std::floor((input.lens()[3] - lengths[1]) /
+                                                      static_cast<float>(stride[1]))) +
+                                1)),
+                    }};
+        }
+        else
+        {
+            MIGRAPHX_THROW("Invalid padding mode");
+        }
+    }
 };
 
 struct leaky_relu
@@ -614,13 +649,13 @@ struct pad
 {
     std::vector<int64_t> pads;
     float value = 0.0f;
-    enum padding_mode_t
+    enum pad_op_mode_t
     {
         constant_pad,
         reflect_pad,
         edge_pad
     };
-    padding_mode_t mode = constant_pad;
+    pad_op_mode_t mode = constant_pad;
 
     template <class Self, class F>
     static auto reflect(Self& self, F f)

src/onnx/onnx.cpp

@@ -215,25 +215,31 @@ struct onnx_parser
     parse_conv(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
     {
         op::convolution op;
+        auto l0 = args[0];
         if(contains(attributes, "pads"))
         {
             if(contains(attributes, "auto_pad"))
             {
                 MIGRAPHX_THROW("auto_pad and padding cannot be specified simultaneously");
             }
-            std::vector<std::size_t> padding(4);
-            copy(attributes["pads"].ints(), padding.begin());
+            std::vector<std::int64_t> padding;
+            copy(attributes["pads"].ints(), std::back_inserter(padding));
             if(padding.size() != 4)
             {
                 MIGRAPHX_THROW("padding should have 4 values");
             }
             if(padding[0] != padding[2] || padding[1] != padding[3])
             {
-                MIGRAPHX_THROW("migraphx does not support asymetric padding");
+                // insert zeros for pad op (args[0] has 4 dims)
+                padding = {0, 0, padding[0], padding[1], 0, 0, padding[2], padding[3]};
+                l0      = prog.add_instruction(op::pad{padding}, l0);
             }
+            else
+            {
                 op.padding[0] = padding[0];
                 op.padding[1] = padding[1];
             }
+        }
         if(contains(attributes, "strides"))
         {
             copy(attributes["strides"].ints(), op.stride.begin());
@@ -252,7 +258,7 @@ struct onnx_parser
 
             if(s.find("SAME") != std::string::npos)
             {
-                op.padding_mode = op::convolution::same;
+                op.padding_mode = op::padding_mode_t::same;
             }
         }
         if(contains(attributes, "group"))
@@ -266,7 +272,7 @@ struct onnx_parser
             auto l2       = prog.add_instruction(op::broadcast{axis, l1->get_shape()}, args[2]);
             return prog.add_instruction(op::add{}, l1, l2);
         }
-        return prog.add_instruction(op, args);
+        return prog.add_instruction(op, l0, args[1]);
     }
 
     instruction_ref parse_pooling(const std::string& name,
@@ -274,6 +280,7 @@ struct onnx_parser
                                   std::vector<instruction_ref> args)
     {
         op::pooling op{ends_with(name, "MaxPool") ? "max" : "average"};
+        auto l0 = args[0];
         if(starts_with(name, "Global"))
         {
             auto lens  = args.front()->get_shape().lens();
@@ -281,19 +288,24 @@ struct onnx_parser
         }
         if(contains(attributes, "pads"))
         {
-            std::vector<std::size_t> padding(4);
-            copy(attributes["pads"].ints(), padding.begin());
+            std::vector<std::int64_t> padding;
+            copy(attributes["pads"].ints(), std::back_inserter(padding));
             if(padding.size() != 4)
             {
                 MIGRAPHX_THROW("padding should have 4 values");
             }
             if(padding[0] != padding[2] || padding[1] != padding[3])
             {
-                MIGRAPHX_THROW("migraphx does not support asymetric padding");
+                // insert zeros for pad op (args[0] has 4 dims)
+                padding = {0, 0, padding[0], padding[1], 0, 0, padding[2], padding[3]};
+                l0      = prog.add_instruction(op::pad{padding}, l0);
             }
+            else
+            {
                 op.padding[0] = padding[0];
                 op.padding[1] = padding[1];
             }
+        }
         if(contains(attributes, "strides"))
         {
             copy(attributes["strides"].ints(), op.stride.begin());
@@ -305,13 +317,14 @@ struct onnx_parser
         if(contains(attributes, "auto_pad"))
         {
             auto s = attributes["auto_pad"].s();
-            if(to_upper(s) != "NOTSET")
+            if(s.find("SAME_UPPER") == std::string::npos)
             {
-                MIGRAPHX_THROW("auto_pad is not supported for pooling");
+                MIGRAPHX_THROW("auto_pad only supports SAME_UPPER for pooling");
             }
+            op.padding_mode = op::padding_mode_t::same;
         }
 
-        return prog.add_instruction(op, std::move(args));
+        return prog.add_instruction(op, l0);
     }
 
     instruction_ref

test/gpu/miopen.cpp

@@ -953,6 +953,22 @@ struct test_pad
     }
 };
 
+struct test_pooling_autopad
+{
+    migraphx::program create_program() const
+    {
+        migraphx::program p;
+        migraphx::shape s0{migraphx::shape::float_type, {1, 3, 63, 63}};
+        auto l0 = p.add_parameter("x", s0);
+        migraphx::op::pooling op{"max"};
+        op.padding_mode = migraphx::op::padding_mode_t::same;
+        op.lengths      = {2, 2};
+        op.stride       = {2, 2};
+        p.add_instruction(op, l0);
+        return p;
+    }
+};
+
 struct test_gather
 {
     migraphx::program create_program() const
@@ -1070,6 +1086,7 @@ struct test_conv_bn_relu_pooling2
 
 int main()
 {
+    verify_program<test_pooling_autopad>();
     verify_program<test_abs>();
     verify_program<test_concat>();
     verify_program<test_concat2>();