未验证 提交 db4ae6e6 编写于 作者: Z zhanglirong1999 提交者: GitHub

[OneDNN] cpu_quantize_pass fix (#56303)

* add cpu quantize pass unit test
上级 1f987a75
...@@ -548,16 +548,6 @@ void CPUQuantizePass::QuantizeConv(Graph* graph, ...@@ -548,16 +548,6 @@ void CPUQuantizePass::QuantizeConv(Graph* graph,
conv_op->Op()->SetAttr("force_fp32_output", true); conv_op->Op()->SetAttr("force_fp32_output", true);
} }
// change threshold in bounded ReLu
if (conv_op->Op()->GetAttrIfExists<std::string>("fuse_activation") ==
"relu6") {
float scale_out =
PADDLE_GET_CONST(float, conv_op->Op()->GetAttr("Scale_out"));
float threshold =
PADDLE_GET_CONST(float, conv_op->Op()->GetAttr("fuse_alpha"));
conv_op->Op()->SetAttr("fuse_alpha", scale_out * threshold);
}
++quantize_conv_count; ++quantize_conv_count;
}; };
......
...@@ -10,6 +10,18 @@ cc_test( ...@@ -10,6 +10,18 @@ cc_test(
device_context device_context
enforce enforce
generated_static_op) generated_static_op)
cc_test(
test_mkldnn_cpu_quantize_pass
SRCS test_mkldnn_cpu_quantize_pass.cc
DEPS executor
op_registry
activation_op
conv_activation_mkldnn_fuse_pass
cpu_quantize_placement_pass
cpu_quantize_pass
phi
scope
device_context)
set(TEST_MKLDNN_CACHING_DEPS set(TEST_MKLDNN_CACHING_DEPS
op_registry op_registry
......
/* Copyright (c) 2023 PaddlePaddle 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. */
#include <glog/logging.h>
#include <gtest/gtest.h>
#include <fstream>
#include <iostream>
#include <unordered_map>
#include "paddle/fluid/framework/ir/fuse_pass_base.h"
#include "paddle/fluid/framework/ir/graph.h"
#include "paddle/fluid/framework/ir/pass.h"
#include "paddle/fluid/framework/ir/pass_tester_helper.h"
#include "paddle/fluid/framework/op_desc.h"
#include "paddle/fluid/framework/program_desc.h"
#include "paddle/fluid/platform/enforce.h"
using std::pair;
using std::string;
using std::unordered_map;
DEFINE_bool(enable_mkldnn, true, "Enable MKLDNN");
namespace paddle {
namespace pass {
using VarQuantScale =
std::unordered_map<std::string, std::pair<bool, phi::DenseTensor>>;
static float const SCALE = 2.f;
const std::vector<std::string> PreGraphPasses({
"conv_activation_mkldnn_fuse_pass",
"cpu_quantize_placement_pass",
"cpu_quantize_pass",
});
TEST(cpuQuantizePass, ConvReLU6) {
paddle::framework::ProgramDesc prog;
auto* block = prog.MutableBlock(0);
auto* conv2d_op = block->AppendOp();
conv2d_op->SetType("conv2d");
conv2d_op->SetInput("Input", {"conv2d-X"});
conv2d_op->SetInput("Filter", {"conv2d-Y"});
conv2d_op->SetOutput("Output", {"conv2d-Out"});
const std::vector<int> strides({1, 1});
const std::vector<int> paddings({1, 1});
const std::vector<int> dilations({1, 1});
const int groups = 1;
conv2d_op->SetAttr("strides", strides);
conv2d_op->SetAttr("paddings", paddings);
conv2d_op->SetAttr("dilations", dilations);
conv2d_op->SetAttr("groups", groups);
auto* relu6_op = block->AppendOp();
relu6_op->SetType("relu6");
relu6_op->SetAttr("threshold", 6.f);
relu6_op->SetInput("X", {"conv2d-Out"});
relu6_op->SetOutput("Out", {"relu-Out"});
auto place = phi::CPUPlace();
VarQuantScale* scales = new VarQuantScale();
phi::DenseTensor scale_input_tensor;
phi::DenseTensor scale_weight_tensor;
scale_input_tensor.Resize({1});
scale_weight_tensor.Resize({1});
auto* ptr_scale_input = scale_input_tensor.mutable_data<double>(place);
auto* ptr_scale_weight = scale_weight_tensor.mutable_data<double>(place);
ptr_scale_input[0] = SCALE;
ptr_scale_weight[0] = SCALE;
(*scales)["conv2d-X"] = std::make_pair(false, std::move(scale_input_tensor));
(*scales)["conv2d-Y"] = std::make_pair(false, std::move(scale_weight_tensor));
paddle::framework::Scope scope;
std::unique_ptr<paddle::framework::ir::Graph> graph(
new paddle::framework::ir::Graph(prog));
(graph)->SetNotOwned(paddle::framework::ir::kParamScopeAttr, &scope);
for (const auto& pass : PreGraphPasses) {
auto pass_ = paddle::framework::ir::PassRegistry::Instance().Get(pass);
if (pass == "cpu_quantize_pass") {
pass_->Set("quant_var_scales", scales);
}
graph.reset(pass_->Apply(graph.release()));
}
int fused_conv2d_num = 0;
for (auto* node : graph->Nodes()) {
if (node->IsOp() && node->Op() && node->Op()->Type() == "fused_conv2d") {
CHECK_EQ(node->Op()->GetAttrIfExists<float>("fuse_beta"), 6)
<< "Attr fuse_beta must equal to 6.";
fused_conv2d_num++;
}
}
CHECK_GT(fused_conv2d_num, 0) << "Graph must contain fused_conv2d";
}
} // namespace pass
} // namespace paddle
USE_PASS(conv_activation_mkldnn_fuse_pass);
USE_PASS(cpu_quantize_placement_pass);
USE_PASS(cpu_quantize_pass);
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册