Rv master

.drone.yml

@@ -0,0 +1,81 @@
+---
+kind: pipeline
+name: TengineRV64
+platform:
+  os: linux
+  arch: amd64 
+
+steps:
+  - name: build
+    image: ubuntu20.04:qemu
+    commands:
+      - PATH=$PATH:/home/riscv/bin cmake -DCMAKE_TOOLCHAIN_FILE=toolchains/rv64-c906.toolchain.cmake -DCMAKE_EXPORT_COMPILE_COMMANDS=ON -DCMAKE_BUILD_TYPE=RELEASE -DTENGINE_BUILD_TESTS=ON -DTENGINE_COVERAGE=ON -B build
+      - PATH=$PATH:/home/riscv/bin cmake --build build -- -j`cat /proc/cpuinfo | grep 'processor' | wc -l` VERBOSE=1 
+  - name: test ops 
+    image: ubuntu20.04:qemu
+    commands:
+      - cd build
+      - export QEMU_CMD='qemu-riscv64 -cpu rv64,v=true -E TG_DEBUG_TIME=1 -L /home/riscv/sysroot'
+      - ../tests/test_rv64_ops.sh
+  - name: test models 
+    image: ubuntu20.04:qemu
+    environment:
+      DATA_SERVER_URL:
+        from_secret: DATA_SERVER_URL
+    commands:
+      - cd build
+      - wget -nv $${DATA_SERVER_URL}/tengine_model_zoo/ci_data/models.tar.gz
+      - wget -nv $${DATA_SERVER_URL}/tengine_model_zoo/ci_data/images.tar.gz
+      - wget -nv $${DATA_SERVER_URL}/tengine_model_zoo/ci_data/data_x86.tar.gz
+      - mkdir models images data
+      - tar zxvf models.tar.gz -C models
+      - tar zxvf images.tar.gz -C images
+      - tar zxvf data_x86.tar.gz -C data
+      - export QEMU_CMD='qemu-riscv64 -cpu rv64,v=true -E TG_DEBUG_TIME=1 -L /home/riscv/sysroot'
+      - ../tests/test_rv64_models.sh
+    when:
+      branch:
+      - master
+  - name: code coverage 
+    image: ubuntu20.04:qemu
+    commands:
+      - cd build
+      - apt update && apt install lcov -y
+      - lcov --gcov-tool /home/riscv/bin/riscv64-unknown-linux-gnu-gcov --capture --directory . --output-file $${DRONE_REPO_NAME}.info
+      - genhtml --branch-coverage -o ../codecov $${DRONE_REPO_NAME}.info 
+  - name: scp files
+    image: appleboy/drone-scp
+    settings:
+      host: conleylee.com
+      username:
+        from_secret: download_host_user
+      password: 
+        from_secret: download_host_passwd
+      port: 38000
+      target: /home/lee/codecov/${DRONE_REPO_NAME}/${DRONE_BUILD_NUMBER}/${DRONE_COMMIT_SHA}
+      strip_components: 1
+      source: codecov/*
+  - name: upload_to_codecov 
+    image: robertstettner/drone-codecov:latest 
+    settings:
+      token:
+        from_secret: CODECOV_TOKEN
+      files: 
+        - build/${DRONE_REPO_NAME}.info
+      flags:
+        - model_test
+  - name: notify
+    image: ubuntu20.04:drone_script 
+    environment:
+      MATTERMOST_TOKEN:
+        from_secret: MATTERMOST_TOKEN
+      GITEA_API_TOKEN:
+        from_secret: gitea_api_token
+    commands:
+      - 'export DRONE_SCRIPT_DOWNLOAD_LINK=https://download.conleylee.com/scripts/drone_bot.py'
+      - 'export DRONE_CODECOV_LINK=https://codecov.conleylee.com/$${DRONE_REPO_NAME}/$${DRONE_BUILD_NUMBER}/$${DRONE_COMMIT_SHA}'
+      - 'wget $${DRONE_SCRIPT_DOWNLOAD_LINK}'
+      - pip3 install mattermostdriver
+      - python3 `basename $${DRONE_SCRIPT_DOWNLOAD_LINK}` 
+    when:
+      status: [success, failure]

CMakeLists.txt

@@ -35,18 +35,6 @@ ENDIF()
 # Enable the languages which in use
 ENABLE_LANGUAGE (C CXX)
 
-IF (CMAKE_TOOLCHAIN_FILE)
-    SET (LIBRARY_OUTPUT_PATH_ROOT ${CMAKE_BINARY_DIR} CACHE PATH "root for library output, set this to change where android libs are compiled to")
-
-    # get absolute path, but get_filename_component ABSOLUTE only refer with source dir, so find_file here :(
-    GET_FILENAME_COMPONENT (CMAKE_TOOLCHAIN_FILE_NAME ${CMAKE_TOOLCHAIN_FILE} NAME)
-    FIND_FILE (CMAKE_TOOLCHAIN_FILE ${CMAKE_TOOLCHAIN_FILE_NAME} PATHS ${CMAKE_SOURCE_DIR} NO_DEFAULT_PATH)
-    MESSAGE (STATUS "Using CMake tool chain file ${CMAKE_TOOLCHAIN_FILE}")
-ENDIF()
-
-IF (NOT CMAKE_BUILD_TYPE)
-    SET (CMAKE_BUILD_TYPE release CACHE STRING "Choose the type of build" FORCE)
-ENDIF()
 
 # Module options
 OPTION (TENGINE_BUILD_BENCHMARK             "Build benchmark"                           ON)
@@ -92,7 +80,23 @@ OPTION (TENGINE_ENABLE_ALL_SYMBOL           "All symbol visible"
 OPTION (TENGINE_ENABLE_MODEL_CACHE          "NPU kernel cache file option"              OFF)
 
 # Online report
-OPTION (TENGINE_ONLINE_REPORT               "online report"                             ON)
+OPTION (TENGINE_ONLINE_REPORT               "online report"                             OFF)
+
+OPTION (TENGINE_RV64_RVV_C906               "build for c906"                            OFF)
+OPTION (TENGINE_COVERAGE                    "build with coverage info"                  OFF)
+
+IF (CMAKE_TOOLCHAIN_FILE)
+    SET (LIBRARY_OUTPUT_PATH_ROOT ${CMAKE_BINARY_DIR} CACHE PATH "root for library output, set this to change where android libs are compiled to")
+
+    # get absolute path, but get_filename_component ABSOLUTE only refer with source dir, so find_file here :(
+    GET_FILENAME_COMPONENT (CMAKE_TOOLCHAIN_FILE_NAME ${CMAKE_TOOLCHAIN_FILE} NAME)
+    FIND_FILE (CMAKE_TOOLCHAIN_FILE ${CMAKE_TOOLCHAIN_FILE_NAME} PATHS ${CMAKE_SOURCE_DIR} NO_DEFAULT_PATH)
+    MESSAGE (STATUS "Using CMake tool chain file ${CMAKE_TOOLCHAIN_FILE}")
+ENDIF()
+
+IF (NOT CMAKE_BUILD_TYPE)
+    SET (CMAKE_BUILD_TYPE release CACHE STRING "Choose the type of build" FORCE)
+ENDIF()
 
 # Do check list
 INCLUDE ("${CMAKE_CURRENT_SOURCE_DIR}/cmake/check.cmake")

README.md

@@ -7,11 +7,9 @@
 
 # Tengine
 
-[![GitHub license](http://OAID.github.io/pics/apache_2.0.svg)](./LICENSE)
-[![GitHub Workflow Status](https://img.shields.io/github/actions/workflow/status/OAID/Tengine/build-and-test.yml?branch=tengine-lite)](https://github.com/OAID/Tengine/actions)
-[![Test Status](https://img.shields.io/travis/OAID/Tengine/tengine-lite?label=test)](https://travis-ci.org/OAID/Tengine)
-[![codecov](https://codecov.io/gh/OAID/Tengine/branch/tengine-lite/graph/badge.svg?token=kz9NcQPRrk)](https://codecov.io/gh/OAID/Tengine)
-[![Language grade: C/C++](https://img.shields.io/lgtm/grade/cpp/g/OAID/Tengine.svg?logo=lgtm&logoWidth=18)](https://lgtm.com/projects/g/OAID/Tengine/context:cpp)
+[![License](https://img.shields.io/badge/license-Apache_2.0-blue)](./LICENSE)
+[![Build Status](https://drone.conleylee.com/api/badges/conley/Tengine/status.svg?ref=refs/heads/master)](https://drone.conleylee.com/conley/Tengine)
+[![codecov](https://codecov.io/gh/ComingToy/Tengine/graph/badge.svg?token=KVOX0LW1NJ)](https://codecov.io/gh/ComingToy/Tengine)
 
 
 ## 简介

README_EN.md

@@ -7,13 +7,9 @@ English | [简体中文](./README.md)
 
 # Tengine
 
-[![GitHub license](http://OAID.github.io/pics/apache_2.0.svg)](./LICENSE)
-[![Build Status](https://img.shields.io/github/workflow/status/OAID/Tengine/Tengine-Lite-Actions/tengine-lite)](https://github.com/OAID/Tengine/actions?query=workflow%3ATengine-Lite-Actions)
-[![Build Status](https://img.shields.io/github/workflow/status/OAID/Tengine-Convert-Tools/Tengine-Convert-Tools-Actions?label=tools%20build)](https://github.com/OAID/Tengine-Convert-Tools/actions?query=workflow%3ATengine-Convert-Tools-Actions)
-[![Test Status](https://img.shields.io/travis/OAID/Tengine/tengine-lite?label=test)](https://travis-ci.org/OAID/Tengine)
-[![codecov](https://codecov.io/gh/OAID/Tengine/branch/tengine-lite/graph/badge.svg?token=kz9NcQPRrk)](https://codecov.io/gh/OAID/Tengine)
-[![Language grade: C/C++](https://img.shields.io/lgtm/grade/cpp/g/OAID/Tengine.svg?logo=lgtm&logoWidth=18)](https://lgtm.com/projects/g/OAID/Tengine/context:cpp)
-
+[![License](https://img.shields.io/badge/license-Apache_2.0-blue)](./LICENSE)
+[![Build Status](https://drone.conleylee.com/api/badges/conley/Tengine/status.svg?ref=refs/heads/master)](https://drone.conleylee.com/conley/Tengine)
+[![codecov](https://codecov.io/gh/ComingToy/Tengine/graph/badge.svg?token=KVOX0LW1NJ)](https://codecov.io/gh/ComingToy/Tengine)
 
 ## Introduction
 

examples/CMakeLists.txt

@@ -62,7 +62,9 @@ TENGINE_EXAMPLE (tm_efficientdet_uint8       tm_efficientdet_uint8.c)
 TENGINE_EXAMPLE (tm_mobilenet_ssd            tm_mobilenet_ssd.c)
 TENGINE_EXAMPLE (tm_mobilenet_ssd_uint8      tm_mobilenet_ssd_uint8.cpp)
 TENGINE_EXAMPLE (tm_retinaface               tm_retinaface.cpp)
+TENGINE_EXAMPLE (tm_retinaface_vulkan        tm_retinaface_vulkan.cpp)
 TENGINE_EXAMPLE (tm_landmark                 tm_landmark.cpp)
+TENGINE_EXAMPLE (tm_landmark_vulkan          tm_landmark_vulkan.cpp)
 TENGINE_EXAMPLE (tm_landmark_uint8           tm_landmark_uint8.cpp)
 TENGINE_EXAMPLE (tm_mobilefacenet            tm_mobilefacenet.cpp)
 TENGINE_EXAMPLE (tm_mobilefacenet_uint8      tm_mobilefacenet_uint8.cpp)

examples/tm_landmark_vulkan.cpp

@@ -0,0 +1,206 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you 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.
+ */
+
+/*
+ * Copyright (c) 2020, OPEN AI LAB
+ * Author: [email protected]
+ */
+
+#include <iostream>
+#include <functional>
+
+#include "common.h"
+#include "tengine/c_api.h"
+#include "tengine_operations.h"
+
+#define DEFAULT_REPEAT_COUNT 1
+#define DEFAULT_THREAD_COUNT 1
+
+void get_input_fp32_data(const char* image_file, float* input_data, int img_h, int img_w, float* mean, float* scale)
+{
+    image img = imread_process(image_file, img_w, img_h, mean, scale);
+
+    float* image_data = (float*)img.data;
+
+    for (int i = 0; i < img_w * img_h * 3; i++)
+        input_data[i] = image_data[i];
+
+    free_image(img);
+}
+
+void show_usage()
+{
+    fprintf(stderr, "[Usage]:  [-h]\n    [-m model_file] [-i image_file] [-r repeat_count] [-t thread_count]\n");
+}
+
+int main(int argc, char* argv[])
+{
+    int repeat_count = DEFAULT_REPEAT_COUNT;
+    int num_thread = DEFAULT_THREAD_COUNT;
+    char* model_file = nullptr;
+    char* image_file = nullptr;
+    int img_h = 144;
+    int img_w = 144;
+    float mean[3] = {128.f, 128.f, 128.f};
+    float scale[3] = {0.0039, 0.0039, 0.0039};
+
+    int res;
+    while ((res = getopt(argc, argv, "m:i:r:t:h:")) != -1)
+    {
+        switch (res)
+        {
+        case 'm':
+            model_file = optarg;
+            break;
+        case 'i':
+            image_file = optarg;
+            break;
+        case 'r':
+            repeat_count = atoi(optarg);
+            break;
+        case 't':
+            num_thread = atoi(optarg);
+            break;
+        case 'h':
+            show_usage();
+            return 0;
+        default:
+            break;
+        }
+    }
+
+    /* check files */
+    if (model_file == nullptr)
+    {
+        fprintf(stderr, "Error: Tengine model file not specified!\n");
+        show_usage();
+        return -1;
+    }
+
+    if (image_file == nullptr)
+    {
+        fprintf(stderr, "Error: Image file not specified!\n");
+        show_usage();
+        return -1;
+    }
+
+    if (!check_file_exist(model_file) || !check_file_exist(image_file))
+        return -1;
+
+    /* set runtime options */
+    struct options opt;
+    opt.num_thread = num_thread;
+    opt.cluster = TENGINE_CLUSTER_ALL;
+    opt.precision = TENGINE_MODE_FP32;
+    opt.affinity = 0;
+
+    /* inital tengine */
+    init_tengine();
+    fprintf(stderr, "tengine-lite library version: %s\n", get_tengine_version());
+
+    /* create graph, load tengine model xxx.tmfile */
+    context_t vk_context = create_context("VK", 1);
+    add_context_device(vk_context, "VK");
+    graph_t graph = create_graph(vk_context, "tengine", model_file);
+    set_graph_device(graph, "VK");
+    if (graph == nullptr)
+    {
+        std::cout << "Create graph0 failed\n";
+        return -1;
+    }
+
+    /* set the input shape to initial the graph, and prerun graph to infer shape */
+    int img_size = img_h * img_w * 3;
+    int dims[] = {1, 3, img_h, img_w}; // nchw
+    float* input_data = (float*)malloc(img_size * sizeof(float));
+
+    tensor_t input_tensor = get_graph_input_tensor(graph, 0, 0);
+    if (input_tensor == nullptr)
+    {
+        fprintf(stderr, "Get input tensor failed\n");
+        return -1;
+    }
+
+    if (set_tensor_shape(input_tensor, dims, 4) < 0)
+    {
+        fprintf(stderr, "Set input tensor shape failed\n");
+        return -1;
+    }
+
+    if (set_tensor_buffer(input_tensor, input_data, img_size * sizeof(float)) < 0)
+    {
+        fprintf(stderr, "Set input tensor buffer failed\n");
+        return -1;
+    }
+
+    /* prerun graph, set work options(num_thread, cluster, precision) */
+    if (prerun_graph_multithread(graph, opt) < 0)
+    {
+        fprintf(stderr, "Prerun multithread graph failed.\n");
+        return -1;
+    }
+
+    /* prepare process input data, set the data mem to input tensor */
+    get_input_fp32_data(image_file, input_data, img_h, img_w, mean, scale);
+
+    /* run graph */
+    double min_time = DBL_MAX;
+    double max_time = DBL_MIN;
+    double total_time = 0.;
+    for (int i = 0; i < repeat_count; i++)
+    {
+        double start = get_current_time();
+        if (run_graph(graph, 1) < 0)
+        {
+            fprintf(stderr, "Run graph failed\n");
+            return -1;
+        }
+        double end = get_current_time();
+        double cur = end - start;
+        total_time += cur;
+        if (min_time > cur)
+            min_time = cur;
+        if (max_time < cur)
+            max_time = cur;
+    }
+    printf("Repeat [%d] min %.3f ms, max %.3f ms, avg %.3f ms\n", repeat_count, min_time, max_time,
+           total_time / repeat_count);
+
+    /* get output tensor */
+    tensor_t output_tensor = get_graph_output_tensor(graph, 0, 0);
+
+    float* data = (float*)(get_tensor_buffer(output_tensor));
+    int data_size = get_tensor_buffer_size(output_tensor) / sizeof(float);
+
+    image img_out = imread(image_file);
+    for (int i = 0; i < data_size / 2; i++)
+    {
+        int x = (int)(data[2 * i] * (float)img_out.w / 144.f);
+        int y = (int)(data[2 * i + 1] * (float)img_out.h / 144.f);
+        draw_circle(img_out, x, y, 2, 0, 255, 0);
+    }
+
+    save_image(img_out, "landmark_out");
+
+    postrun_graph(graph);
+    destroy_graph(graph);
+    release_tengine();
+
+    return 0;
+}