qlinearadd operator (#2188)

src/onnx/broadcast_qdq.cpp

@@ -0,0 +1,76 @@
+/*
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2015-2023 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include <migraphx/onnx/broadcast_qdq.hpp>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace onnx {
+
+// This method is to prep for quantizelinear or dequantizelinear operation for
+// either the broadcasting of weight-scale or zero-points of qlinearadd operator
+// outputs: operator op (inputs x, broadcasted: scale (float) & zero_pt (8-bit))
+instruction_ref bcast_qdq_instr(const std::string& op_name,
+                                instruction_ref x_in,
+                                instruction_ref arg_fscale,
+                                instruction_ref arg_z_pt,
+                                const onnx_parser::node_info& info)
+{
+    auto in_lens = x_in->get_shape().lens();
+
+    // prep 1: broadcast scale. it can come as a scalar or a 1-D tensor.
+    instruction_ref bcast_scale;
+    if(arg_fscale->get_shape().elements() > 1)
+        bcast_scale = info.add_instruction(
+            migraphx::make_op("broadcast", {{"axis", 0}, {"out_lens", in_lens}}), arg_fscale);
+    else
+        bcast_scale = info.add_instruction(
+            migraphx::make_op("multibroadcast", {{"out_lens", in_lens}}), arg_fscale);
+
+    // prep 2: broadcast zero point. it can come as a scalar or a 1-D tensor.
+    instruction_ref bcast_zero_pt;
+    if(arg_z_pt->get_shape().elements() > 1)
+        bcast_zero_pt = info.add_instruction(
+            migraphx::make_op("broadcast", {{"axis", 0}, {"out_lens", in_lens}}), arg_z_pt);
+    else
+        bcast_zero_pt = info.add_instruction(
+            migraphx::make_op("multibroadcast", {{"out_lens", in_lens}}), arg_z_pt);
+
+    // op_name is either quantizelinear or dequantizelinear:
+    return info.add_instruction(migraphx::make_op(op_name), x_in, bcast_scale, bcast_zero_pt);
+}
+
+// Multibroadcast a scaler..
+instruction_ref bcast_scalar_instr(const migraphx::shape& shape_out,
+                                   instruction_ref arg_in,
+                                   const onnx_parser::node_info& info)
+{
+    auto bcast_instr_out = info.add_instruction(
+        migraphx::make_op("multibroadcast", {{"out_lens", shape_out.lens()}}), arg_in);
+    return bcast_instr_out;
+}
+
+} // namespace onnx
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx

src/onnx/include/migraphx/onnx/broadcast_qdq.hpp

@@ -0,0 +1,56 @@
+/*
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2015-2023 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#ifndef MIGRAPHX_GUARD_AMDMIGRAPHX_ONNX_BROADCAST_QDQ_HPP
+#define MIGRAPHX_GUARD_AMDMIGRAPHX_ONNX_BROADCAST_QDQ_HPP
+
+#include <string>
+
+#include <migraphx/onnx/op_parser.hpp>
+#include <migraphx/make_op.hpp>
+#include <migraphx/instruction.hpp>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace onnx {
+
+// This method is to prep for quantizelinear or dequantizelinear operation for
+// either the broadcasting of weight-scale or zero-points of qlinearadd operator
+// outputs: operator op (inputs x, broadcasted: scale (float) & zero_pt (8-bit))
+instruction_ref bcast_qdq_instr(const std::string& op_name,
+                                instruction_ref x_in,
+                                instruction_ref arg_fscale,
+                                instruction_ref arg_z_pt,
+                                const onnx_parser::node_info& info);
+
+// Multibroadcast a scaler..
+instruction_ref bcast_scalar_instr(const migraphx::shape& shape_out,
+                                   instruction_ref arg_in,
+                                   const onnx_parser::node_info& info);
+
+} // namespace onnx
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/onnx/parse_qlinearadd.cpp

@@ -0,0 +1,154 @@
+/*
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2015-2023 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include <migraphx/onnx/op_parser.hpp>
+#include <migraphx/ranges.hpp>
+#include <migraphx/common.hpp>
+#include <migraphx/make_op.hpp>
+#include <migraphx/onnx/checks.hpp>
+#include <migraphx/onnx/broadcast_qdq.hpp>
+#include <migraphx/instruction.hpp>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace onnx {
+
+/*
+ *********************************************************************************
+ *  Reference: see QLinearAdd in                                                 *
+ *  https://github.com/microsoft/onnxruntime/blob/main/docs/ContribOperators.md  *
+ *********************************************************************************
+
+  com.microsoft.QLinearAdd
+  Performs element-wise binary addition on 8 bit data types (with Numpy-style broadcasting support).
+
+  C = (A_scale * (A - A_zero_point) + B_scale * (B - B_zero_point))/C_scale + C_zero_point
+
+  Version
+  This version of the operator has been available since version 1 of the 'com.microsoft' operator
+  set.
+
+  Inputs (7 - 8)
+  A : T
+  First operand.
+
+  A_scale : tensor(float)
+  Input A's scale. It's a scalar, which means a per-tensor/layer quantization.
+
+  A_zero_point (optional) : T
+  Input A zero point. Default value is 0 if it's not specified. It's a scalar, which means a
+  per-tensor/layer quantization.
+
+  B : T
+  Second operand.
+
+  B_scale : tensor(float)
+  Input B's scale. It's a scalar, which means a per-tensor/layer quantization.
+
+  B_zero_point (optional) : T
+  Input B zero point. Default value is 0 if it's not specified. It's a scalar, which means a
+  per-tensor/layer quantization.
+
+  C_scale : tensor(float)
+  Output scale. It's a scalar, which means a per-tensor/layer quantization.
+
+  C_zero_point (optional) : T
+
+  Output zero point. Default value is 0 if it's not specified. It's a scalar, which means a
+  per-tensor/layer quantization.
+
+  Outputs
+  C : T
+  Result, has same element type as two inputs
+
+  Type Constraints
+  T : tensor(uint8), tensor(int8)
+  Constrain input and output types to 8 bit signed and unsigned tensors.
+
+*/
+
+struct parse_qlinearadd : op_parser<parse_qlinearadd>
+{
+    std::vector<op_desc> operators() const { return {{"QLinearAdd"}}; }
+
+    // basic type checking for QLinearAdd Operator
+    void check_inputs(const std::vector<instruction_ref>& args) const
+    {
+        if(args.size() < 7)
+            MIGRAPHX_THROW("QLINEARADD: missing inputs");
+
+        const auto& in_a = args[0];
+        const auto& in_b = args[3];
+
+        auto sh_a = in_a->get_shape();
+        auto sh_b = in_b->get_shape();
+
+        auto type_a = sh_a.type();
+        auto type_b = sh_b.type();
+        if(type_a != migraphx::shape::int8_type and type_a != migraphx::shape::uint8_type)
+            MIGRAPHX_THROW("QLINEARADD: unsupported input type");
+        if(type_b != migraphx::shape::int8_type and type_b != migraphx::shape::uint8_type)
+            MIGRAPHX_THROW("QLINEARADD: unsupported input type");
+        if(type_a != type_b)
+            MIGRAPHX_THROW("QLINEARADD: mismatched input types");
+    }
+
+    instruction_ref parse(const op_desc& /* opd */,
+                          const onnx_parser& /*parser*/,
+                          const onnx_parser::node_info& info,
+                          const std::vector<instruction_ref>& args) const
+    {
+        check_inputs(args);
+
+        // A
+        const auto& in_a         = args[0];
+        const auto& in_scale_a   = args[1];
+        const auto& in_zero_pt_a = args[2];
+
+        auto dquant_a = bcast_qdq_instr("dequantizelinear", in_a, in_scale_a, in_zero_pt_a, info);
+
+        // B
+        const auto& in_b         = args[3];
+        const auto& in_scale_b   = args[4];
+        const auto& in_zero_pt_b = args[5];
+        auto dquant_b = bcast_qdq_instr("dequantizelinear", in_b, in_scale_b, in_zero_pt_b, info);
+
+        // C = A + B
+        auto out_c = info.add_common_op("add", dquant_a, dquant_b);
+
+        const auto& in_scale_c = args[6];
+
+        // zero_pt for C is supplied as the last optional argument..
+        if(args.size() == 8)
+            return (bcast_qdq_instr("quantizelinear", out_c, in_scale_c, args[7], info));
+
+        // if no zero_pt: just broadcast the scale..
+        auto bcast_scale_c = bcast_scalar_instr(out_c->get_shape(), in_scale_c, info);
+        return (info.add_instruction(migraphx::make_op("quantizelinear"), out_c, bcast_scale_c));
+    }
+};
+
+} // namespace onnx
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx

test/onnx/gen_onnx.py

@@ -5096,6 +5096,61 @@ def prelu_brcst_test():
     return ([node], [arg0, arg1], [arg_out])
 
 
+@onnx_test()
+def qlinearadd_test():
+    a = helper.make_tensor_value_info('A', TensorProto.UINT8, [64])
+    sc_a = helper.make_tensor('A_scale', TensorProto.FLOAT, [], [0.05])
+    zero_pt_a = helper.make_tensor('A_zero_point', TensorProto.UINT8, [], [0])
+
+    b = helper.make_tensor_value_info('B', TensorProto.UINT8, [64])
+    sc_b = helper.make_tensor('B_scale', TensorProto.FLOAT, [], [0.05])
+    zero_pt_b = helper.make_tensor('B_zero_point', TensorProto.UINT8, [],
+                                   [128])
+
+    sc_c = helper.make_tensor('C_scale', TensorProto.FLOAT, [], [0.05])
+    zero_pt_c = helper.make_tensor('C_zero_point', TensorProto.UINT8, [], [64])
+
+    c = helper.make_tensor_value_info('C', TensorProto.UINT8, [64])
+
+    node = onnx.helper.make_node(
+        'QLinearAdd',
+        inputs=[
+            'A', 'A_scale', 'A_zero_point', 'B', 'B_scale', 'B_zero_point',
+            'C_scale', 'C_zero_point'
+        ],
+        outputs=['C'],
+    )
+    return ([node], [a, b], [c],
+            [sc_a, zero_pt_a, sc_b, zero_pt_b, sc_c, zero_pt_c])
+
+
+@onnx_test()
+def qlinearadd_bcast_test():
+    a = helper.make_tensor_value_info('A', TensorProto.INT8, [64])
+    sc_a = helper.make_tensor('A_scale', TensorProto.FLOAT, [], [0.05])
+    zero_pt_a = helper.make_tensor('A_zero_point', TensorProto.INT8, [], [0])
+
+    b = helper.make_tensor_value_info('B', TensorProto.INT8, [1, 1, 64])
+    sc_b = helper.make_tensor('B_scale', TensorProto.FLOAT, [], [0.05])
+    zero_pt_b = helper.make_tensor('B_zero_point', TensorProto.INT8, [], [32])
+
+    sc_c = helper.make_tensor('C_scale', TensorProto.FLOAT, [], [0.05])
+    zero_pt_c = helper.make_tensor('C_zero_point', TensorProto.INT8, [], [-64])
+
+    c = helper.make_tensor_value_info('C', TensorProto.INT8, [1, 1, 64])
+
+    node = onnx.helper.make_node(
+        'QLinearAdd',
+        inputs=[
+            'A', 'A_scale', 'A_zero_point', 'B', 'B_scale', 'B_zero_point',
+            'C_scale', 'C_zero_point'
+        ],
+        outputs=['C'],
+    )
+    return ([node], [a, b], [c],
+            [sc_a, zero_pt_a, sc_b, zero_pt_b, sc_c, zero_pt_c])
+
+
 @onnx_test()
 def quantizelinear_test():
     arg0 = helper.make_tensor_value_info('0', TensorProto.FLOAT, [5])