[xpu] update resnet50 ut and add googlenet, vgg19 uts (#4277)

lite/CMakeLists.txt

@@ -51,11 +51,18 @@ if (WITH_TESTING)
         lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL} "GoogleNet_inference.tar.gz")
         lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL} "mobilenet_v1.tar.gz")
         lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL} "mobilenet_v2_relu.tar.gz")
-        lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL} "resnet50.tar.gz")
         lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL} "inception_v4_simple.tar.gz")
         lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL} "step_rnn.tar.gz")
-        lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL} "bert.tar.gz")
-        lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL} "ernie.tar.gz")
+
+        set(LITE_URL_FOR_UNITTESTS "http://paddle-inference-dist.bj.bcebos.com/PaddleLite/models_and_data_for_unittests")
+        # models
+        lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL_FOR_UNITTESTS} "resnet50.tar.gz")
+        lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL_FOR_UNITTESTS} "bert.tar.gz")
+        lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL_FOR_UNITTESTS} "ernie.tar.gz")
+        lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL_FOR_UNITTESTS} "GoogLeNet.tar.gz")
+        lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL_FOR_UNITTESTS} "VGG19.tar.gz")
+        # data
+        lite_download_and_uncompress(${LITE_MODEL_DIR} ${LITE_URL_FOR_UNITTESTS} "ILSVRC2012_small.tar.gz")
     endif()
 endif()
 

lite/tests/api/CMakeLists.txt

@@ -8,33 +8,26 @@ if(LITE_WITH_ARM)
     endif()
 endif()
 
-if(LITE_WITH_XPU AND NOT LITE_WITH_XTCL)
-    lite_cc_test(test_resnet50_fp32_xpu SRCS test_resnet50_fp32_xpu.cc
-      DEPS mir_passes lite_api_test_helper paddle_api_full paddle_api_light gflags utils
-      ${ops} ${host_kernels} ${x86_kernels} ${xpu_kernels}
-      ARGS --model_dir=${LITE_MODEL_DIR}/resnet50)
-    lite_cc_test(test_ernie_fp32_xpu SRCS test_ernie_fp32_xpu.cc
-      DEPS mir_passes lite_api_test_helper paddle_api_full paddle_api_light gflags utils
-      ${ops} ${host_kernels} ${x86_kernels} ${xpu_kernels}
-      ARGS --model_dir=${LITE_MODEL_DIR}/ernie)
-    lite_cc_test(test_bert_fp32_xpu SRCS test_bert_fp32_xpu.cc
+function(xpu_x86_without_xtcl_test TARGET MODEL DATA)
+    lite_cc_test(${TARGET} SRCS ${TARGET}.cc
         DEPS mir_passes lite_api_test_helper paddle_api_full paddle_api_light gflags utils
         ${ops} ${host_kernels} ${x86_kernels} ${xpu_kernels}
-      ARGS --model_dir=${LITE_MODEL_DIR}/bert)
+        ARGS --model_dir=${LITE_MODEL_DIR}/${MODEL}
+             --data_dir=${LITE_MODEL_DIR}/${DATA})
     if(WITH_TESTING)
-        add_dependencies(test_resnet50_fp32_xpu extern_lite_download_resnet50_tar_gz)
-        add_dependencies(test_ernie_fp32_xpu extern_lite_download_ernie_tar_gz)
-        add_dependencies(test_bert_fp32_xpu extern_lite_download_bert_tar_gz)
+        add_dependencies(${TARGET} extern_lite_download_${MODEL}_tar_gz)
+        if(NOT ${DATA} STREQUAL "")
+            add_dependencies(${TARGET} extern_lite_download_${DATA}_tar_gz)
+        endif()
     endif()
-    # TODO(miaotianxiang): enable later
-    #lite_cc_test(test_fpr_fp32_xpu SRCS test_fpr_fp32_xpu.cc
-      #DEPS mir_passes lite_api_test_helper paddle_api_full paddle_api_light gflags utils
-      #${ops} ${host_kernels} ${x86_kernels} ${xpu_kernels}
-      #ARGS --model_dir=${LITE_MODEL_DIR}/resnet50)
-    #lite_cc_test(test_mmdnn_fp32_xpu SRCS test_mmdnn_fp32_xpu.cc
-      #DEPS mir_passes lite_api_test_helper paddle_api_full paddle_api_light gflags utils
-      #${ops} ${host_kernels} ${x86_kernels} ${xpu_kernels}
-      #ARGS --model_dir=${LITE_MODEL_DIR}/resnet50)
+endfunction()
+
+if(LITE_WITH_XPU AND NOT LITE_WITH_XTCL)
+    xpu_x86_without_xtcl_test(test_resnet50_fp32_xpu resnet50 ILSVRC2012_small)
+    xpu_x86_without_xtcl_test(test_googlenet_fp32_xpu GoogLeNet ILSVRC2012_small)
+    xpu_x86_without_xtcl_test(test_vgg19_fp32_xpu VGG19 ILSVRC2012_small)
+    xpu_x86_without_xtcl_test(test_ernie_fp32_xpu ernie "")
+    xpu_x86_without_xtcl_test(test_bert_fp32_xpu bert "")
 endif()
 
 if(LITE_WITH_RKNPU)

lite/tests/api/ILSVRC2012_utility.h

+// Copyright (c) 2020 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.
+
+#pragma once
+#include <algorithm>
+#include <iostream>
+#include <string>
+#include <vector>
+#include "lite/utils/cp_logging.h"
+#include "lite/utils/io.h"
+#include "lite/utils/string.h"
+
+namespace paddle {
+namespace lite {
+
+template <class T = float>
+std::vector<std::vector<T>> ReadRawData(
+    const std::string& raw_data_dir,
+    const std::vector<int>& input_shape = {1, 3, 224, 224},
+    int iteration = 100) {
+  std::vector<std::vector<T>> raw_data;
+
+  int image_size = 1;
+  for (size_t i = 1; i < input_shape.size(); i++) {
+    image_size *= input_shape[i];
+  }
+  int input_size = image_size * input_shape[0];
+
+  for (int i = 0; i < iteration; i++) {
+    std::vector<T> one_iter_raw_data;
+    one_iter_raw_data.resize(input_size);
+    T* data = &(one_iter_raw_data.at(0));
+    for (int j = 0; j < input_shape[0]; j++) {
+      std::string raw_data_file_dir =
+          raw_data_dir + std::string("/") +
+          std::to_string(i * input_shape[0] + j + 1);
+      std::ifstream fin(raw_data_file_dir, std::ios::in | std::ios::binary);
+      CHECK(fin.is_open()) << "failed to open file " << raw_data_file_dir;
+      fin.seekg(0, std::ios::end);
+      int file_size = fin.tellg();
+      fin.seekg(0, std::ios::beg);
+      CHECK_EQ(file_size, image_size * sizeof(T) / sizeof(char));
+      fin.read(reinterpret_cast<char*>(data), file_size);
+      fin.close();
+      data += image_size;
+    }
+    raw_data.emplace_back(one_iter_raw_data);
+  }
+
+  return raw_data;
+}
+
+float CalOutAccuracy(const std::vector<std::vector<float>>& out_rets,
+                     const std::string& labels_dir) {
+  std::vector<int> labels;
+  std::vector<int> out_top1;
+  int right_num = 0;
+
+  auto label_lines = ReadLines(labels_dir);
+  for (size_t i = 0; i < out_rets.size(); i++) {
+    int label = std::stoi(Split(label_lines[i], " ")[1]);
+
+    auto out = out_rets[i];
+    auto largest = std::max_element(out.begin(), out.end());
+    int out_top1 = std::distance(out.begin(), largest);
+
+    right_num += (out_top1 == label);
+  }
+
+  return static_cast<float>(right_num) / static_cast<float>(out_rets.size());
+}
+
+}  // namespace lite
+}  // namespace paddle

lite/tests/api/test_googlenet_fp32_xpu.cc

+// Copyright (c) 2019 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 <gflags/gflags.h>
+#include <gtest/gtest.h>
+#include <vector>
+#include "lite/api/lite_api_test_helper.h"
+#include "lite/api/paddle_api.h"
+#include "lite/api/paddle_use_kernels.h"
+#include "lite/api/paddle_use_ops.h"
+#include "lite/api/paddle_use_passes.h"
+#include "lite/api/test_helper.h"
+#include "lite/tests/api/ILSVRC2012_utility.h"
+#include "lite/utils/cp_logging.h"
+
+DEFINE_string(data_dir, "", "data dir");
+DEFINE_int32(iteration, 100, "iteration times to run");
+DEFINE_int32(batch, 1, "batch of image");
+DEFINE_int32(channel, 3, "image channel");
+
+namespace paddle {
+namespace lite {
+
+TEST(GoogLeNet, test_googlenet_fp32_xpu) {
+  lite_api::CxxConfig config;
+  config.set_model_dir(FLAGS_model_dir);
+  config.set_valid_places({lite_api::Place{TARGET(kXPU), PRECISION(kFloat)},
+                           lite_api::Place{TARGET(kX86), PRECISION(kFloat)},
+                           lite_api::Place{TARGET(kHost), PRECISION(kFloat)}});
+  config.set_xpu_workspace_l3_size_per_thread();
+  auto predictor = lite_api::CreatePaddlePredictor(config);
+
+  std::string raw_data_dir = FLAGS_data_dir + std::string("/raw_data");
+  std::vector<int> input_shape{
+      FLAGS_batch, FLAGS_channel, FLAGS_im_width, FLAGS_im_height};
+  auto raw_data = ReadRawData(raw_data_dir, input_shape, FLAGS_iteration);
+
+  int input_size = 1;
+  for (auto i : input_shape) {
+    input_size *= i;
+  }
+
+  for (int i = 0; i < FLAGS_warmup; ++i) {
+    auto input_tensor = predictor->GetInput(0);
+    input_tensor->Resize(
+        std::vector<int64_t>(input_shape.begin(), input_shape.end()));
+    auto* data = input_tensor->mutable_data<float>();
+    for (int j = 0; j < input_size; j++) {
+      data[j] = 0.f;
+    }
+    predictor->Run();
+  }
+
+  std::vector<std::vector<float>> out_rets;
+  out_rets.resize(FLAGS_iteration);
+  double cost_time = 0;
+  for (size_t i = 0; i < raw_data.size(); ++i) {
+    auto input_tensor = predictor->GetInput(0);
+    input_tensor->Resize(
+        std::vector<int64_t>(input_shape.begin(), input_shape.end()));
+    auto* data = input_tensor->mutable_data<float>();
+    memcpy(data, raw_data[i].data(), sizeof(float) * input_size);
+
+    double start = GetCurrentUS();
+    predictor->Run();
+    cost_time += GetCurrentUS() - start;
+
+    auto output_tensor = predictor->GetOutput(0);
+    auto output_shape = output_tensor->shape();
+    auto output_data = output_tensor->data<float>();
+    ASSERT_EQ(output_shape.size(), 2UL);
+    ASSERT_EQ(output_shape[0], 1);
+    ASSERT_EQ(output_shape[1], 1000);
+
+    int output_size = output_shape[0] * output_shape[1];
+    out_rets[i].resize(output_size);
+    memcpy(&(out_rets[i].at(0)), output_data, sizeof(float) * output_size);
+  }
+
+  LOG(INFO) << "================== Speed Report ===================";
+  LOG(INFO) << "Model: " << FLAGS_model_dir << ", threads num " << FLAGS_threads
+            << ", warmup: " << FLAGS_warmup << ", batch: " << FLAGS_batch
+            << ", iteration: " << FLAGS_iteration << ", spend "
+            << cost_time / FLAGS_iteration / 1000.0 << " ms in average.";
+
+  std::string labels_dir = FLAGS_data_dir + std::string("/labels.txt");
+  float out_accuracy = CalOutAccuracy(out_rets, labels_dir);
+  ASSERT_GT(out_accuracy, 0.57f);
+}
+
+}  // namespace lite
+}  // namespace paddle

lite/tests/api/test_resnet50_fp32_xpu.cc

@@ -21,8 +21,14 @@
 #include "lite/api/paddle_use_ops.h"
 #include "lite/api/paddle_use_passes.h"
 #include "lite/api/test_helper.h"
+#include "lite/tests/api/ILSVRC2012_utility.h"
 #include "lite/utils/cp_logging.h"
 
+DEFINE_string(data_dir, "", "data dir");
+DEFINE_int32(iteration, 100, "iteration times to run");
+DEFINE_int32(batch, 1, "batch of image");
+DEFINE_int32(channel, 3, "image channel");
+
 namespace paddle {
 namespace lite {
 
@@ -35,52 +41,62 @@ TEST(Resnet50, test_resnet50_fp32_xpu) {
   config.set_xpu_workspace_l3_size_per_thread();
   auto predictor = lite_api::CreatePaddlePredictor(config);
 
-  auto input_tensor = predictor->GetInput(0);
-  std::vector<int64_t> input_shape{1, 3, 224, 224};
-  input_tensor->Resize(input_shape);
-  auto* data = input_tensor->mutable_data<float>();
-  int input_num = 1;
-  for (size_t i = 0; i < input_shape.size(); ++i) {
-    input_num *= input_shape[i];
-  }
-  for (int i = 0; i < input_num; i++) {
-    data[i] = 1;
+  std::string raw_data_dir = FLAGS_data_dir + std::string("/raw_data");
+  std::vector<int> input_shape{
+      FLAGS_batch, FLAGS_channel, FLAGS_im_width, FLAGS_im_height};
+  auto raw_data = ReadRawData(raw_data_dir, input_shape, FLAGS_iteration);
+
+  int input_size = 1;
+  for (auto i : input_shape) {
+    input_size *= i;
   }
 
   for (int i = 0; i < FLAGS_warmup; ++i) {
+    auto input_tensor = predictor->GetInput(0);
+    input_tensor->Resize(
+        std::vector<int64_t>(input_shape.begin(), input_shape.end()));
+    auto* data = input_tensor->mutable_data<float>();
+    for (int j = 0; j < input_size; j++) {
+      data[j] = 0.f;
+    }
     predictor->Run();
   }
 
-  auto start = GetCurrentUS();
-  for (int i = 0; i < FLAGS_repeats; ++i) {
+  std::vector<std::vector<float>> out_rets;
+  out_rets.resize(FLAGS_iteration);
+  double cost_time = 0;
+  for (size_t i = 0; i < raw_data.size(); ++i) {
+    auto input_tensor = predictor->GetInput(0);
+    input_tensor->Resize(
+        std::vector<int64_t>(input_shape.begin(), input_shape.end()));
+    auto* data = input_tensor->mutable_data<float>();
+    memcpy(data, raw_data[i].data(), sizeof(float) * input_size);
+
+    double start = GetCurrentUS();
     predictor->Run();
+    cost_time += GetCurrentUS() - start;
+
+    auto output_tensor = predictor->GetOutput(0);
+    auto output_shape = output_tensor->shape();
+    auto output_data = output_tensor->data<float>();
+    ASSERT_EQ(output_shape.size(), 2UL);
+    ASSERT_EQ(output_shape[0], 1);
+    ASSERT_EQ(output_shape[1], 1000);
+
+    int output_size = output_shape[0] * output_shape[1];
+    out_rets[i].resize(output_size);
+    memcpy(&(out_rets[i].at(0)), output_data, sizeof(float) * output_size);
   }
 
   LOG(INFO) << "================== Speed Report ===================";
   LOG(INFO) << "Model: " << FLAGS_model_dir << ", threads num " << FLAGS_threads
-            << ", warmup: " << FLAGS_warmup << ", repeats: " << FLAGS_repeats
-            << ", spend " << (GetCurrentUS() - start) / FLAGS_repeats / 1000.0
-            << " ms in average.";
-
-  std::vector<std::vector<float>> results;
-  results.emplace_back(std::vector<float>(
-      {0.000268651, 0.000174053, 0.000213181, 0.000396771, 0.000591516,
-       0.00018169,  0.000289721, 0.000855934, 0.000732185, 9.2055e-05,
-       0.000220664, 0.00235289,  0.00571265,  0.00357688,  0.00129667,
-       0.000465392, 0.000143775, 0.000211628, 0.000617144, 0.000265033}));
-  auto out = predictor->GetOutput(0);
-  ASSERT_EQ(out->shape().size(), 2);
-  ASSERT_EQ(out->shape()[0], 1);
-  ASSERT_EQ(out->shape()[1], 1000);
+            << ", warmup: " << FLAGS_warmup << ", batch: " << FLAGS_batch
+            << ", iteration: " << FLAGS_iteration << ", spend "
+            << cost_time / FLAGS_iteration / 1000.0 << " ms in average.";
 
-  int step = 50;
-  for (size_t i = 0; i < results.size(); ++i) {
-    for (size_t j = 0; j < results[i].size(); ++j) {
-      EXPECT_NEAR(out->data<float>()[j * step + (out->shape()[1] * i)],
-                  results[i][j],
-                  1e-5);
-    }
-  }
+  std::string labels_dir = FLAGS_data_dir + std::string("/labels.txt");
+  float out_accuracy = CalOutAccuracy(out_rets, labels_dir);
+  ASSERT_GT(out_accuracy, 0.6f);
 }
 
 }  // namespace lite

lite/tests/api/test_vgg19_fp32_xpu.cc

+// Copyright (c) 2019 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 <gflags/gflags.h>
+#include <gtest/gtest.h>
+#include <vector>
+#include "lite/api/lite_api_test_helper.h"
+#include "lite/api/paddle_api.h"
+#include "lite/api/paddle_use_kernels.h"
+#include "lite/api/paddle_use_ops.h"
+#include "lite/api/paddle_use_passes.h"
+#include "lite/api/test_helper.h"
+#include "lite/tests/api/ILSVRC2012_utility.h"
+#include "lite/utils/cp_logging.h"
+
+DEFINE_string(data_dir, "", "data dir");
+DEFINE_int32(iteration, 100, "iteration times to run");
+DEFINE_int32(batch, 1, "batch of image");
+DEFINE_int32(channel, 3, "image channel");
+
+namespace paddle {
+namespace lite {
+
+TEST(VGG19, test_vgg19_fp32_xpu) {
+  lite_api::CxxConfig config;
+  config.set_model_dir(FLAGS_model_dir);
+  config.set_valid_places({lite_api::Place{TARGET(kXPU), PRECISION(kFloat)},
+                           lite_api::Place{TARGET(kX86), PRECISION(kFloat)},
+                           lite_api::Place{TARGET(kHost), PRECISION(kFloat)}});
+  config.set_xpu_workspace_l3_size_per_thread();
+  auto predictor = lite_api::CreatePaddlePredictor(config);
+
+  std::string raw_data_dir = FLAGS_data_dir + std::string("/raw_data");
+  std::vector<int> input_shape{
+      FLAGS_batch, FLAGS_channel, FLAGS_im_width, FLAGS_im_height};
+  auto raw_data = ReadRawData(raw_data_dir, input_shape, FLAGS_iteration);
+
+  int input_size = 1;
+  for (auto i : input_shape) {
+    input_size *= i;
+  }
+
+  for (int i = 0; i < FLAGS_warmup; ++i) {
+    auto input_tensor = predictor->GetInput(0);
+    input_tensor->Resize(
+        std::vector<int64_t>(input_shape.begin(), input_shape.end()));
+    auto* data = input_tensor->mutable_data<float>();
+    for (int j = 0; j < input_size; j++) {
+      data[j] = 0.f;
+    }
+    predictor->Run();
+  }
+
+  std::vector<std::vector<float>> out_rets;
+  out_rets.resize(FLAGS_iteration);
+  double cost_time = 0;
+  for (size_t i = 0; i < raw_data.size(); ++i) {
+    auto input_tensor = predictor->GetInput(0);
+    input_tensor->Resize(
+        std::vector<int64_t>(input_shape.begin(), input_shape.end()));
+    auto* data = input_tensor->mutable_data<float>();
+    memcpy(data, raw_data[i].data(), sizeof(float) * input_size);
+
+    double start = GetCurrentUS();
+    predictor->Run();
+    cost_time += GetCurrentUS() - start;
+
+    auto output_tensor = predictor->GetOutput(0);
+    auto output_shape = output_tensor->shape();
+    auto output_data = output_tensor->data<float>();
+    ASSERT_EQ(output_shape.size(), 2UL);
+    ASSERT_EQ(output_shape[0], 1);
+    ASSERT_EQ(output_shape[1], 1000);
+
+    int output_size = output_shape[0] * output_shape[1];
+    out_rets[i].resize(output_size);
+    memcpy(&(out_rets[i].at(0)), output_data, sizeof(float) * output_size);
+  }
+
+  LOG(INFO) << "================== Speed Report ===================";
+  LOG(INFO) << "Model: " << FLAGS_model_dir << ", threads num " << FLAGS_threads
+            << ", warmup: " << FLAGS_warmup << ", batch: " << FLAGS_batch
+            << ", iteration: " << FLAGS_iteration << ", spend "
+            << cost_time / FLAGS_iteration / 1000.0 << " ms in average.";
+
+  std::string labels_dir = FLAGS_data_dir + std::string("/labels.txt");
+  float out_accuracy = CalOutAccuracy(out_rets, labels_dir);
+  ASSERT_GT(out_accuracy, 0.56f);
+}
+
+}  // namespace lite
+}  // namespace paddle