conv_bn_fuse_pass.cc 15.7 KB
Newer Older
S
Sylwester Fraczek 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
// Copyright (c) 2018 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 "paddle/fluid/framework/ir/conv_bn_fuse_pass.h"
16
#include <algorithm>
S
Sylwester Fraczek 已提交
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53
#include <functional>
#include <string>
#include <vector>
#include "paddle/fluid/framework/lod_tensor.h"
#include "paddle/fluid/operators/math/cpu_vec.h"
#include "paddle/fluid/platform/enforce.h"

namespace paddle {
namespace framework {
namespace ir {

#define GET_CONV_BN_NODES(pattern_name)                                      \
  /* OPERATORS */                                                            \
  GET_IR_NODE_FROM_SUBGRAPH(conv, conv, pattern_name);                       \
  GET_IR_NODE_FROM_SUBGRAPH(batch_norm, batch_norm, pattern_name);           \
  /* CONV inputs */                                                          \
  GET_IR_NODE_FROM_SUBGRAPH(conv_weight, conv_weight, pattern_name);         \
  /* CONV outputs */                                                         \
  GET_IR_NODE_FROM_SUBGRAPH(conv_out, conv_out, pattern_name);               \
  /* BN inputs */                                                            \
  GET_IR_NODE_FROM_SUBGRAPH(bn_scale, bn_scale, pattern_name);               \
  GET_IR_NODE_FROM_SUBGRAPH(bn_bias, bn_bias, pattern_name);                 \
  GET_IR_NODE_FROM_SUBGRAPH(bn_mean, bn_mean, pattern_name);                 \
  GET_IR_NODE_FROM_SUBGRAPH(bn_variance, bn_variance, pattern_name);         \
  /* BN outputs */                                                           \
  GET_IR_NODE_FROM_SUBGRAPH(bn_out, bn_out, pattern_name); /* Out */         \
  GET_IR_NODE_FROM_SUBGRAPH(bn_mean_out, bn_mean_out, pattern_name);         \
  GET_IR_NODE_FROM_SUBGRAPH(bn_variance_out, bn_variance_out, pattern_name); \
  GET_IR_NODE_FROM_SUBGRAPH(bn_saved_mean, bn_saved_mean, pattern_name);     \
  GET_IR_NODE_FROM_SUBGRAPH(bn_saved_variance, bn_saved_variance, pattern_name)

void recompute_bias_and_weights(const Scope* scope,
                                ir::Node* conv_weight,            //
                                const ir::Node& bn_scale,         //
                                const LoDTensor& bn_bias_tensor,  //
                                const ir::Node& bn_mean,          //
                                const ir::Node& bn_variance,      //
54
                                LoDTensor* eltwise_y_in_tensor,   //
55
                                float epsilon, const std::string& conv_type) {
56 57 58 59 60 61 62
  using EigenVectorArrayMap =
      Eigen::Map<Eigen::Array<float, Eigen::Dynamic, 1>>;
  using ConstEigenVectorArrayMap =
      Eigen::Map<const Eigen::Array<float, Eigen::Dynamic, 1>>;
  using EigenMatrixArrayMap = Eigen::Map<
      Eigen::Array<float, Eigen::Dynamic, Eigen::Dynamic, Eigen::RowMajor>>;

S
Sylwester Fraczek 已提交
63
  // Re-compute bias of conv2d from BN
64 65 66 67 68 69
  PADDLE_ENFORCE_EQ(
      eltwise_y_in_tensor->dims(), bn_bias_tensor.dims(),
      platform::errors::InvalidArgument("Tensor elementwise y(%d) and batch "
                                        "norm bias(%d) must have same dims.",
                                        eltwise_y_in_tensor->dims().size(),
                                        bn_bias_tensor.dims().size()));
S
Sylwester Fraczek 已提交
70 71 72 73 74 75

  auto* scale_tensor = scope->FindVar(bn_scale.Name())->GetMutable<LoDTensor>();
  auto* variance_tensor =
      scope->FindVar(bn_variance.Name())->GetMutable<LoDTensor>();
  auto* mean_tensor = scope->FindVar(bn_mean.Name())->GetMutable<LoDTensor>();

76 77 78 79 80 81 82 83 84
  ConstEigenVectorArrayMap scale_array(scale_tensor->data<float>(),
                                       scale_tensor->numel(), 1);
  EigenVectorArrayMap variance_array(
      variance_tensor->mutable_data<float>(platform::CPUPlace()),
      variance_tensor->numel(), 1);
  ConstEigenVectorArrayMap mean_array(mean_tensor->data<float>(),
                                      mean_tensor->numel(), 1);
  ConstEigenVectorArrayMap bn_bias_array(bn_bias_tensor.data<float>(),
                                         bn_bias_tensor.numel(), 1);
S
Sylwester Fraczek 已提交
85

86 87 88 89 90 91 92 93
  // variance will not be used anymore, so make it std_array and then tmp_array
  variance_array += epsilon;
  variance_array = variance_array.sqrt();
  variance_array = scale_array / variance_array;

  EigenVectorArrayMap eltwise_y_in_array(
      eltwise_y_in_tensor->mutable_data<float>(platform::CPUPlace()),
      eltwise_y_in_tensor->numel(), 1);
94

95 96
  eltwise_y_in_array =
      ((eltwise_y_in_array - mean_array) * variance_array) + bn_bias_array;
S
Sylwester Fraczek 已提交
97 98

  // Re-compute weight of conv2d from BN
99 100
  auto* weights = scope->FindVar(conv_weight->Name())->GetMutable<LoDTensor>();
  auto weights_shape = weights->dims();
101 102 103 104 105 106 107 108 109 110 111 112 113 114
  auto weights_data = weights->mutable_data<float>(platform::CPUPlace());

  // ConvTranspose weights are in IOHW format
  if (conv_type == "conv2d_transpose") {
    int kernel_size = weights_shape[2] * weights_shape[3];
    for (int i = 0; i < weights->numel();) {
      for (int j = 0; j < weights_shape[1]; ++j) {
        for (int k = 0; k < kernel_size; ++k, ++i) {
          weights_data[i] *= variance_array[j];
        }
      }
    }
  } else {
    auto weights_shape_2d = flatten_to_2d(weights_shape, 1);
115

116 117
    EigenMatrixArrayMap weights_array_2d(weights_data, weights_shape_2d[0],
                                         weights_shape_2d[1]);
118

119 120
    weights_array_2d.colwise() *= variance_array;
  }
S
Sylwester Fraczek 已提交
121 122
}

123
void ConvBNFusePass::ApplyImpl(ir::Graph* graph) const {
124 125
  PADDLE_ENFORCE_NOT_NULL(
      graph, platform::errors::InvalidArgument("Graph cannot be nullptr."));
126
  FusePassBase::Init(name_scope_, graph);
S
Sylwester Fraczek 已提交
127 128

  auto* scope = param_scope();
129 130
  PADDLE_ENFORCE_NOT_NULL(
      scope, platform::errors::InvalidArgument("Scope cannot be nullptr."));
S
Sylwester Fraczek 已提交
131 132 133 134 135 136

  GraphPatternDetector gpd;
  auto* conv_input =
      gpd.mutable_pattern()
          ->NewNode(patterns::PDNodeName(name_scope_, "conv_input"))
          ->AsInput()
137
          ->assert_is_op_input(conv_type(), "Input");
S
Sylwester Fraczek 已提交
138
  patterns::ConvBN conv_bn_pattern(gpd.mutable_pattern(), name_scope_);
139
  conv_bn_pattern(conv_input, conv_type(), false /*with_eltwise_add*/);
S
Sylwester Fraczek 已提交
140 141 142 143

  int found_conv_bn_count = 0;
  auto handler = [&](const GraphPatternDetector::subgraph_t& subgraph,
                     Graph* g) {
144
    VLOG(4) << "handle " + conv_type() + "BN fuse";
S
Sylwester Fraczek 已提交
145 146 147 148

    // conv, batch_norm,
    // conv_weight, conv_out,
    // bn_scale, bn_bias, bn_mean, bn_variance,
W
Wojciech Uss 已提交
149 150
    // bn_out, bn_mean_out, bn_variance_out, bn_saved_mean,
    // bn_saved_variance
S
Sylwester Fraczek 已提交
151 152
    GET_CONV_BN_NODES(conv_bn_pattern);

W
Wojciech Uss 已提交
153 154 155
    // check if fuse can be done and if MKL-DNN should be used
    FuseOptions fuse_option = FindFuseOption(*conv, *batch_norm);
    if (fuse_option == DO_NOT_FUSE) {
156
      VLOG(3) << "do not perform " + conv_type() + " bn fuse";
W
Wojciech Uss 已提交
157 158 159
      return;
    }

160 161 162 163
    // Get batch norm bias
    auto* bn_bias_tensor =
        scope->FindVar(bn_bias->Name())->GetMutable<LoDTensor>();

S
Sylwester Fraczek 已提交
164 165 166
    // Create eltwise_y (conv bias) variable
    VarDesc eltwise_y_in_desc(
        patterns::PDNodeName(name_scope_, "eltwise_y_in"));
167 168 169
    eltwise_y_in_desc.SetShape(framework::vectorize(bn_bias_tensor->dims()));
    eltwise_y_in_desc.SetDataType(bn_bias_tensor->type());
    eltwise_y_in_desc.SetLoDLevel(bn_bias->Var()->GetLoDLevel());
W
Wojciech Uss 已提交
170
    eltwise_y_in_desc.SetPersistable(true);
S
Sylwester Fraczek 已提交
171 172 173 174 175 176 177 178 179 180
    auto* eltwise_y_in_node = g->CreateVarNode(&eltwise_y_in_desc);
    auto* eltwise_y_in_tensor =
        scope->Var(eltwise_y_in_node->Name())->GetMutable<LoDTensor>();

    // Initialize eltwise_y
    eltwise_y_in_tensor->Resize(bn_bias_tensor->dims());
    std::fill_n(eltwise_y_in_tensor->mutable_data<float>(platform::CPUPlace()),
                eltwise_y_in_tensor->numel(), 0.0f);

    // update weights and biases
181 182
    float epsilon =
        BOOST_GET_CONST(float, batch_norm->Op()->GetAttr("epsilon"));
S
Sylwester Fraczek 已提交
183
    recompute_bias_and_weights(scope, conv_weight, *bn_scale, *bn_bias_tensor,
184
                               *bn_mean, *bn_variance, eltwise_y_in_tensor,
185
                               epsilon, conv_type());
S
Sylwester Fraczek 已提交
186

W
Wojciech Uss 已提交
187 188 189 190 191 192 193 194 195
    // with MKL-DNN fuse conv+bn into conv with bias
    // without MKL-DNN fuse conv+bn into conv+elementwise_add
    if (fuse_option == FUSE_MKLDNN) {
      auto input_names = conv->Op()->InputNames();
      bool has_bias = std::find(input_names.begin(), input_names.end(),
                                "Bias") != input_names.end();
      if (has_bias && conv->Op()->Input("Bias").size() > 0) {
        // reuse existing conv bias node
        auto conv_bias_names = conv->Op()->Input("Bias");
196 197 198
        PADDLE_ENFORCE_EQ(
            conv_bias_names.size(), 1UL,
            platform::errors::InvalidArgument("Find input var Bais error."));
W
Wojciech Uss 已提交
199 200
        auto* conv_bias_var = scope->FindVar(conv_bias_names[0]);
        auto* conv_bias_tensor = conv_bias_var->GetMutable<LoDTensor>();
201 202 203 204 205 206 207
        PADDLE_ENFORCE_EQ(
            conv_bias_tensor->dims(), eltwise_y_in_tensor->dims(),
            platform::errors::InvalidArgument(
                "Tensor convolution bias(%d) and elementwise y(%d) "
                "must have same dims.",
                conv_bias_tensor->dims().size(),
                eltwise_y_in_tensor->dims().size()));
W
Wojciech Uss 已提交
208 209 210 211 212 213 214 215 216 217 218 219

        auto eigen_conv_bias = EigenVector<float>::From(*conv_bias_tensor);
        eigen_conv_bias += EigenVector<float>::From(*eltwise_y_in_tensor);
      } else {
        // add new conv_bias node
        conv->Op()->SetInput(
            "Bias", std::vector<std::string>({eltwise_y_in_node->Name()}));
        IR_NODE_LINK_TO(eltwise_y_in_node, conv);
      }
      conv->Op()->SetOutput("Output",
                            std::vector<std::string>({bn_out->Name()}));
      GraphSafeRemoveNodes(
220
          graph,
W
Wojciech Uss 已提交
221 222 223 224 225 226 227 228 229 230 231 232 233 234 235
          {conv_out, bn_scale, bn_bias, bn_mean, bn_variance, batch_norm,
           bn_mean_out, bn_variance_out, bn_saved_mean, bn_saved_variance});

      IR_NODE_LINK_TO(conv, bn_out);
      found_conv_bn_count++;
    } else {  // fuse_option == FUSE_NATIVE
      // create an elementwise add node.
      OpDesc desc;
      desc.SetInput("X", std::vector<std::string>({conv_out->Name()}));
      desc.SetInput("Y", std::vector<std::string>({eltwise_y_in_node->Name()}));
      desc.SetOutput("Out", std::vector<std::string>({bn_out->Name()}));
      desc.SetType("elementwise_add");
      desc.SetAttr("axis", 1);
      auto eltwise_op = g->CreateOpNode(&desc);  // OpDesc will be copied.

236 237 238
      GraphSafeRemoveNodes(graph, {bn_scale, bn_bias, bn_mean, bn_variance,
                                   batch_norm, bn_mean_out, bn_variance_out,
                                   bn_saved_mean, bn_saved_variance});
W
Wojciech Uss 已提交
239 240 241 242 243 244

      IR_NODE_LINK_TO(conv_out, eltwise_op);
      IR_NODE_LINK_TO(eltwise_y_in_node, eltwise_op);
      IR_NODE_LINK_TO(eltwise_op, bn_out);
      found_conv_bn_count++;
    }
S
Sylwester Fraczek 已提交
245 246
  };

247
  gpd(graph, handler);
S
Sylwester Fraczek 已提交
248 249 250 251

  AddStatis(found_conv_bn_count);
}

252
void ConvEltwiseAddBNFusePass::ApplyImpl(ir::Graph* graph) const {
253 254
  PADDLE_ENFORCE_NOT_NULL(
      graph, platform::errors::InvalidArgument("Graph cannot be nullptr."));
255
  FusePassBase::Init(name_scope_, graph);
S
Sylwester Fraczek 已提交
256 257

  auto* scope = param_scope();
258 259
  PADDLE_ENFORCE_NOT_NULL(
      scope, platform::errors::InvalidArgument("Scope cannot be nullptr."));
S
Sylwester Fraczek 已提交
260 261 262 263 264 265

  GraphPatternDetector gpd;
  auto* conv_input =
      gpd.mutable_pattern()
          ->NewNode(patterns::PDNodeName(name_scope_, "conv_input"))
          ->AsInput()
266
          ->assert_is_op_input(conv_type(), "Input");
S
Sylwester Fraczek 已提交
267
  patterns::ConvBN conv_bn_pattern(gpd.mutable_pattern(), name_scope_);
268
  conv_bn_pattern(conv_input, conv_type(), true /*with_eltwise_add*/);
S
Sylwester Fraczek 已提交
269 270 271 272

  int found_conv_bn_count = 0;
  auto handler = [&](const GraphPatternDetector::subgraph_t& subgraph,
                     Graph* g) {
273
    VLOG(4) << "handle " + conv_type() + "BN fuse";
S
Sylwester Fraczek 已提交
274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295

    // conv, batch_norm,
    // conv_weight, conv_out,
    // bn_scale, bn_bias, bn_mean, bn_variance,
    // bn_out, bn_mean_out, bn_variance_out, bn_saved_mean,bn_saved_variance
    GET_CONV_BN_NODES(conv_bn_pattern);
    // OPERATORS
    GET_IR_NODE_FROM_SUBGRAPH(eltwise, eltwise, conv_bn_pattern);
    // BIAS inputs
    GET_IR_NODE_FROM_SUBGRAPH(eltwise_y_in, eltwise_y_in, conv_bn_pattern);
    // BIAS outputs
    GET_IR_NODE_FROM_SUBGRAPH(eltwise_out, eltwise_out, conv_bn_pattern);

    // Get eltwise_y (conv bias) variable
    auto* eltwise_y_in_tensor =
        scope->FindVar(eltwise_y_in->Name())->GetMutable<LoDTensor>();

    // Get batch norm bias
    auto* bn_bias_tensor =
        scope->FindVar(bn_bias->Name())->GetMutable<LoDTensor>();

    // update weights and biases
296 297
    float epsilon =
        BOOST_GET_CONST(float, batch_norm->Op()->GetAttr("epsilon"));
298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339

    // if bias is an input to other ops as well then we cannot overwrite it
    // so we create separate elementwise Y in nodes
    if (eltwise_y_in->outputs.size() > 1) {
      // Make a copy of eltwise Y input tensor
      // Create eltwise_y (conv bias) variable
      VarDesc eltwise_y_in_desc(patterns::PDNodeName(
          name_scope_, "eltwise_y_in" + std::to_string(found_conv_bn_count)));
      eltwise_y_in_desc.SetShape(
          framework::vectorize(eltwise_y_in_tensor->dims()));
      eltwise_y_in_desc.SetDataType(eltwise_y_in_tensor->type());
      eltwise_y_in_desc.SetLoDLevel(eltwise_y_in->Var()->GetLoDLevel());
      eltwise_y_in_desc.SetPersistable(true);
      auto* eltwise_y_in_node = g->CreateVarNode(&eltwise_y_in_desc);
      auto* eltwise_y_in_tensor_ex =
          scope->Var(eltwise_y_in_node->Name())->GetMutable<LoDTensor>();

      // Initialize eltwise_y
      TensorCopy(*eltwise_y_in_tensor, platform::CPUPlace(),
                 eltwise_y_in_tensor_ex);

      recompute_bias_and_weights(scope, conv_weight, *bn_scale, *bn_bias_tensor,
                                 *bn_mean, *bn_variance, eltwise_y_in_tensor_ex,
                                 epsilon, conv_type());
      // Set new var
      eltwise->Op()->RenameInput(eltwise_y_in->Name(),
                                 eltwise_y_in_node->Name());
      // Link new bias node to eltwise
      IR_NODE_LINK_TO(eltwise_y_in_node, eltwise);
      // unlink original bias from eltwise_op
      eltwise_y_in->outputs.erase(
          std::remove_if(eltwise_y_in->outputs.begin(),
                         eltwise_y_in->outputs.end(),
                         [&](Node*& n) {
                           return n->id() == eltwise->id() ? true : false;
                         }),
          eltwise_y_in->outputs.end());
    } else {
      recompute_bias_and_weights(scope, conv_weight, *bn_scale, *bn_bias_tensor,
                                 *bn_mean, *bn_variance, eltwise_y_in_tensor,
                                 epsilon, conv_type());
    }
S
Sylwester Fraczek 已提交
340 341 342 343 344 345

    // Update the elementwise_add node
    eltwise->Op()->SetAttr("axis", 1);
    eltwise->Op()->SetOutput("Out", std::vector<std::string>({bn_out->Name()}));

    GraphSafeRemoveNodes(
346
        graph,
S
Sylwester Fraczek 已提交
347 348 349 350 351 352 353 354
        {bn_scale, bn_bias, bn_mean, bn_variance, batch_norm, bn_mean_out,
         bn_variance_out, bn_saved_mean, bn_saved_variance, eltwise_out});

    IR_NODE_LINK_TO(eltwise, bn_out);

    found_conv_bn_count++;
  };

355
  gpd(graph, handler);
S
Sylwester Fraczek 已提交
356 357 358 359 360 361 362 363 364 365 366

  AddStatis(found_conv_bn_count);
}

}  // namespace ir
}  // namespace framework
}  // namespace paddle

REGISTER_PASS(conv_bn_fuse_pass, paddle::framework::ir::ConvBNFusePass);
REGISTER_PASS(conv_eltwiseadd_bn_fuse_pass,
              paddle::framework::ir::ConvEltwiseAddBNFusePass);
367 368 369 370
REGISTER_PASS(conv_transpose_bn_fuse_pass,
              paddle::framework::ir::ConvTransposeBNFusePass);
REGISTER_PASS(conv_transpose_eltwiseadd_bn_fuse_pass,
              paddle::framework::ir::ConvTransposeEltwiseAddBNFusePass);