快速识别你家的猫猫狗狗，教你用ModelBox开发AI萌宠应用-低调大师

快速识别你家的猫猫狗狗，教你用ModelBox开发AI萌宠应用

2024-06-13 295

本文分享自华为云社区《ModelBox-AI应用开发：动物目标检测【玩转华为云】》，作者：阳光大猫。

一、准备环境

二、应用开发

1. 创建工程

在ModelBox sdk目录下使用create.bat创建yolov7_pet工程

(tensorflow) PS D:\modelbox-win10-x64-1.5.3> .\create.bat -t server -n yolov7_pet 

(tensorflow) D:\modelbox-win10-x64-1.5.3>set BASE_PATH=D:\modelbox-win10-x64-1.5.3\ 

(tensorflow) D:\modelbox-win10-x64-1.5.3>set PATH=D:\modelbox-win10-x64-1.5.3\\python-embed;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3\envs\tensorflow;C:\Users\yanso\miniconda3\envs\tensorflow\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\usr\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Scripts;C:\Users\yanso\miniconda3\envs\tensorflow\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\Library\usr\bin;C:\Users\yanso\miniconda3\Library\bin;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin

(tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONPATH=

(tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONHOME=

(tensorflow) D:\modelbox-win10-x64-1.5.3>python.exe -u D:\modelbox-win10-x64-1.5.3\\create.py -t server -n yolov7_pet
sdk version is modelbox-win10-x64-1.5.3
dos2unix: converting file D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet/graph\modelbox.conf to Unix format...
dos2unix: converting file D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet/graph\yolov7_pet.toml to Unix format...
dos2unix: converting file D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet/bin\mock_task.toml to Unix format...
success: create yolov7_pet in D:\modelbox-win10-x64-1.5.3\workspace

create.bat工具的参数中，-t表示所创建实例的类型，包括server（ModelBox工程）、python（Python功能单元）、c++（C++功能单元）、infer（推理功能单元）等；-n表示所创建实例的名称，开发者自行命名。

2. 创建推理功能单元

在ModelBox sdk目录下使用create.bat创建yolov7_infer推理功能单元

(tensorflow) PS D:\modelbox-win10-x64-1.5.3> .\create.bat -t infer -n yolov7_infer -p yolov7_pet  

(tensorflow) D:\modelbox-win10-x64-1.5.3>set BASE_PATH=D:\modelbox-win10-x64-1.5.3\ 

(tensorflow) D:\modelbox-win10-x64-1.5.3>set PATH=D:\modelbox-win10-x64-1.5.3\\python-embed;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3\envs\tensorflow;C:\Users\yanso\miniconda3\envs\tensorflow\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\usr\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Scripts;C:\Users\yanso\miniconda3\envs\tensorflow\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\Library\usr\bin;C:\Users\yanso\miniconda3\Library\bin;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin 

(tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONPATH= 

(tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONHOME= 

(tensorflow) D:\modelbox-win10-x64-1.5.3>python.exe -u D:\modelbox-win10-x64-1.5.3\\create.py -t infer -n yolov7_infer -p yolov7_pet    
sdk version is modelbox-win10-x64-1.5.3
success: create infer yolov7_infer in D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet/model/yolov7_infer

create.bat工具使用时，-t infer 即表示创建的是推理功能单元；-n xxx_infer 表示创建的功能单元名称为xxx_infer；-p yolov7_infer 表示所创建的功能单元属于yolov7_infer应用。

a. 下载转换好的模型

运行此Notebook下载转换好的ONNX格式模型

b. 修改模型配置文件

模型和配置文件保持在同级目录下

# Copyright (C) 2020 Huawei Technologies Co., Ltd. All rights reserved.

[base]
name = "yolov7_infer"
device = "cpu"
version = "1.0.0"
description = "your description"
entry = "./best.onnx"  # model file path, use relative path
type = "inference" 
virtual_type = "onnx" # inference engine type: win10 now only support onnx
group_type = "Inference"  # flowunit group attribution, do not change

# Input ports description
[input]
[input.input1]  # input port number, Format is input.input[N]
name = "Input"  # input port name
type = "float"  # input port data type ,e.g. float or uint8
device = "cpu"  # input buffer type: cpu, win10 now copy input from cpu

# Output ports description
[output]
[output.output1] # output port number, Format is output.output[N]
name = "Output"  # output port name
type = "float"   # output port data type ,e.g. float or uint8

3. 创建后处理功能单元

在ModelBox sdk目录下使用create.bat创建yolov7_post后处理功能单元

(tensorflow) PS D:\modelbox-win10-x64-1.5.3> .\create.bat -t python -n yolov7_post -p yolov7_pet  

(tensorflow) D:\modelbox-win10-x64-1.5.3>set BASE_PATH=D:\modelbox-win10-x64-1.5.3\ 

(tensorflow) D:\modelbox-win10-x64-1.5.3>set PATH=D:\modelbox-win10-x64-1.5.3\\python-embed;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3\envs\tensorflow;C:\Users\yanso\miniconda3\envs\tensorflow\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\usr\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Scripts;C:\Users\yanso\miniconda3\envs\tensorflow\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\Library\usr\bin;C:\Users\yanso\miniconda3\Library\bin;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin

(tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONPATH=

(tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONHOME=

(tensorflow) D:\modelbox-win10-x64-1.5.3>python.exe -u D:\modelbox-win10-x64-1.5.3\\create.py -t python -n yolov7_post -p yolov7_pet
sdk version is modelbox-win10-x64-1.5.3
success: create python yolov7_post in D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet/etc/flowunit/yolov7_post

a. 修改配置文件

# Copyright (c) Huawei Technologies Co., Ltd. 2022. All rights reserved.

# Basic config
[base]
name = "yolov7_post" # The FlowUnit name
device = "cpu" # The flowunit runs on cpu
version = "1.0.0" # The version of the flowunit
type = "python" # Fixed value, do not change
description = "description" # The description of the flowunit
entry = "yolov7_post@yolov7_postFlowUnit" # Python flowunit entry function
group_type = "Generic"  # flowunit group attribution, change as Input/Output/Image/Generic ...

# Flowunit Type
stream = false # Whether the flowunit is a stream flowunit
condition = false # Whether the flowunit is a condition flowunit
collapse = false # Whether the flowunit is a collapse flowunit
collapse_all = false # Whether the flowunit will collapse all the data
expand = false #  Whether the flowunit is a expand flowunit

# The default Flowunit config
[config]
net_h = 640
net_w = 640
num_classes = 2
conf_threshold = 0.5
iou_threshold = 0.45

# Input ports description
[input]
[input.input1] # Input port number, the format is input.input[N]
name = "in_feat" # Input port name
type = "float" # Input port type

# Output ports description
[output]
[output.output1] # Output port number, the format is output.output[N]
name = "out_data" # Output port name
type = "string" # Output port type

b. 修改逻辑代码

# Copyright (c) Huawei Technologies Co., Ltd. 2022. All rights reserved.

#!/usr/bin/env python
# -*- coding: utf-8 -*-
import _flowunit as modelbox
import numpy as np
import json
import cv2

class yolov7_postFlowUnit(modelbox.FlowUnit):
    # Derived from modelbox.FlowUnit
    def __init__(self):
        super().__init__()

    # Open the flowunit to obtain configuration information
    def open(self, config):
        # 获取功能单元的配置参数
        self.params = {}
        self.params['net_h'] = config.get_int('net_h')
        self.params['net_w'] = config.get_int('net_w')
        self.params['num_classes'] = config.get_int('num_classes')
        self.params['conf_thre'] = config.get_float('conf_threshold')
        self.params['nms_thre'] = config.get_float('iou_threshold')
        self.num_classes = config.get_int('num_classes')

        return modelbox.Status.StatusCode.STATUS_SUCCESS

    # Process the data
    def process(self, data_context):
        # 从DataContext中获取输入输出BufferList对象
        in_feat = data_context.input("in_feat")
        out_data = data_context.output("out_data")

        # yolov7_post process code.
        # 循环处理每一个输入Buffer数据
        for buffer_feat in in_feat:
            # 将输入Buffer转换为numpy对象
            feat_data = np.array(buffer_feat.as_object(), copy=False)
            feat_data = feat_data.reshape((-1, self.num_classes + 5))

            # 业务处理：解码yolov7模型的输出数据，得到检测框，转化为json数据
            bboxes = self.postprocess(feat_data, self.params)
            result = {"det_result": str(bboxes)}
            print(result)

            # 将业务处理返回的结果数据转换为Buffer
            result_str = json.dumps(result)
            out_buffer = modelbox.Buffer(self.get_bind_device(), result_str)

            # 将输出Buffer放入输出BufferList中
            out_data.push_back(out_buffer)

        return modelbox.Status.StatusCode.STATUS_SUCCESS
    
    # model post-processing function
    def postprocess(self, feat_data, params):
        """postprocess for yolo7 model"""
        boxes = []
        class_ids = []
        confidences = []
        for detection in feat_data:
            scores = detection[5:]
            class_id = np.argmax(scores)
            if params['num_classes'] == 1:
                confidence = detection[4]
            else:
                confidence = detection[4] * scores[class_id] 
            if confidence > params['conf_thre'] and detection[4] > params['conf_thre']:
                center_x = detection[0] / params['net_w']
                center_y = detection[1] / params['net_h']
                width = detection[2] / params['net_w']
                height = detection[3] / params['net_h']

                left = center_x - width / 2
                top = center_y - height / 2

                class_ids.append(class_id)
                confidences.append(confidence)
                boxes.append([left, top, width, height])

        # use nms algorithm in opencv
        box_idx = cv2.dnn.NMSBoxes(
            boxes, confidences, params['conf_thre'], params['nms_thre'])

        detections = []
        for i in box_idx:
            boxes[i][0] = max(0.0, boxes[i][0])  # [0, 1]
            boxes[i][1] = max(0.0, boxes[i][1])  # [0, 1]
            boxes[i][2] = min(1.0, boxes[i][0] + boxes[i][2])  # [0, 1]
            boxes[i][3] = min(1.0, boxes[i][1] + boxes[i][3])  # [0, 1]
            dets = np.concatenate(
                [boxes[i], np.array([confidences[i]]), np.array([class_ids[i]])], 0).tolist()
            detections.append(dets)

        return detections

    def close(self):
        # Close the flowunit
        return modelbox.Status()

    def data_pre(self, data_context):
        # Before streaming data starts
        return modelbox.Status()

    def data_post(self, data_context):
        # After streaming data ends
        return modelbox.Status()

    def data_group_pre(self, data_context):
        # Before all streaming data starts
        return modelbox.Status()

    def data_group_post(self, data_context):
        # After all streaming data ends
        return modelbox.Status()

4. 修改流程图

yolov7_pet工程graph目录下存放流程图，默认的流程图yolov7_pet.toml与工程同名，其内容为（以Windows版ModelBox为例）：

# Copyright (C) 2020 Huawei Technologies Co., Ltd. All rights reserved.

[driver]
dir = ["${HILENS_APP_ROOT}/etc/flowunit",
"${HILENS_APP_ROOT}/etc/flowunit/cpp",
"${HILENS_APP_ROOT}/model",
"${HILENS_MB_SDK_PATH}/flowunit"]
skip-default = true
[profile]
profile=false
trace=false
dir="${HILENS_DATA_DIR}/mb_profile"
[graph]
format = "graphviz"
graphconf = """digraph yolov7_pet {
    node [shape=Mrecord]
    queue_size = 4
    batch_size = 1
    input1[type=input,flowunit=input,device=cpu,deviceid=0]

    httpserver_sync_receive[type=flowunit, flowunit=httpserver_sync_receive_v2, device=cpu, deviceid=0, time_out_ms=5000, endpoint="http://0.0.0.0:8083/v1/yolov7_pet", max_requests=100]
    image_decoder[type=flowunit, flowunit=image_decoder, device=cpu, key="image_base64", queue_size=4]
    image_resize[type=flowunit, flowunit=resize, device=cpu, deviceid=0, image_width=640, image_height=640]
    image_transpose[type=flowunit, flowunit=packed_planar_transpose, device=cpu, deviceid=0]
    normalize[type=flowunit flowunit=normalize device=cpu deviceid=0 standard_deviation_inverse="0.0039215686,0.0039215686,0.0039215686"]
    yolov7_infer[type=flowunit, flowunit=yolov7_infer, device=cpu, deviceid=0, batch_size = 1]
    yolov7_post[type=flowunit, flowunit=yolov7_post, device=cpu, deviceid=0]
    httpserver_sync_reply[type=flowunit, flowunit=httpserver_sync_reply_v2, device=cpu, deviceid=0]
    
    input1:input -> httpserver_sync_receive:in_url
    httpserver_sync_receive:out_request_info -> image_decoder:in_encoded_image
    image_decoder:out_image -> image_resize:in_image
    image_resize:out_image -> image_transpose:in_image
    image_transpose:out_image -> normalize:in_data
    normalize:out_data -> yolov7_infer:Input
    yolov7_infer:Output -> yolov7_post:in_feat
    yolov7_post:out_data -> httpserver_sync_reply:in_reply_info
}"""
[flow]
desc = "yolov7_pet run in modelbox-win10-x64"

5. 准备动物图片和测试脚本

a. 动物图片

yolov7_pet工程data目录下存放动物图片文件夹test_imgs

b. 测试脚本

yolov7_pet工程data目录下存放测试脚本test_http.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

# Copyright (c) Huawei Technologies Co., Ltd. 2022. All rights reserved.

import os
import cv2
import json
import base64
import http.client
class HttpConfig:
    '''http调用的参数配置'''
    def __init__(self, host_ip, port, url, img_base64_str):
        self.hostIP = host_ip
        self.Port = port

        self.httpMethod = "POST"
        self.requstURL = url
        self.headerdata = {
            "Content-Type": "application/json"
        }
        self.test_data = {
            "image_base64": img_base64_str
        }
        self.body = json.dumps(self.test_data)
def read_image(img_path):
    '''读取图片数据并转为base64编码的字符串'''
    img_data = cv2.imread(img_path)
    img_str = cv2.imencode('.jpg', img_data)[1].tostring()
    img_bin = base64.b64encode(img_str)
    img_base64_str = str(img_bin, encoding='utf8')
    return img_data, img_base64_str
def decode_car_bboxes(bbox_str, input_shape):
    try:
        labels = [0, 1]  # cat, dog
        bboxes = json.loads(json.loads(bbox_str)['det_result'])
        bboxes = list(filter(lambda x: int(x[5]) in labels, bboxes))
    except Exception as ex:
        print(str(ex))
        return []
    else:
        for bbox in bboxes:
            bbox[0] = int(bbox[0] * input_shape[1])
            bbox[1] = int(bbox[1] * input_shape[0])
            bbox[2] = int(bbox[2] * input_shape[1])
            bbox[3] = int(bbox[3] * input_shape[0])
        return bboxes
def draw_bboxes(img_data, bboxes):
    '''画框'''
    for bbox in bboxes:
        x1, y1, x2, y2, score, label = bbox
        color = (0, 0, 255)
        names = ['cat', 'dog']  
        score = '%.2f' % score
        label = '%s:%s' % (names[int(label)], score)
        cv2.rectangle(img_data, (x1, y1), (x2, y2), color, 2)
        cv2.putText(img_data, label, (x1, y1 - 10), cv2.FONT_HERSHEY_TRIPLEX, 0.5, (0, 255, 0), thickness=1)
    return img_data
def test_image(img_path, ip, port, url):
    '''单张图片测试'''
    img_data, img_base64_str = read_image(img_path)
    http_config = HttpConfig(ip, port, url, img_base64_str)

    conn = http.client.HTTPConnection(host=http_config.hostIP, port=http_config.Port)
    conn.request(method=http_config.httpMethod, url=http_config.requstURL,
                body=http_config.body, headers=http_config.headerdata)

    response = conn.getresponse().read().decode()
    print('response: ', response)

    bboxes = decode_car_bboxes(response, img_data.shape)
    imt_out = draw_bboxes(img_data, bboxes)
    cv2.imwrite('./result-' + os.path.basename(img_path), imt_out)
if __name__ == "__main__":
    port = 8083
    ip = "127.0.0.1"
    url = "/v1/yolov7_pet"
    img_path = "./test.jpg"
    img_folder = './test_imgs'
    file_list = os.listdir(img_folder)
    for img_file in file_list:
        print("\n================ {} ================".format(img_file))
        img_path = os.path.join(img_folder, img_file)
        test_image(img_path, ip, port, url)

三、运行应用

在yolov7_pet工程目录下执行.\bin\main.bat运行应用：

(tensorflow) PS D:\modelbox-win10-x64-1.5.3> cd D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet
(tensorflow) PS D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet> .\bin\main.bat

(tensorflow) D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet>set PATH=D:/modelbox-win10-x64-1.5.3/workspace/yolov7_pet/bin/../../../python-embed;D:/modelbox-win10-x64-1.5.3/workspace/yolov7_pet/bin/../../../modelbox-win10-x64/bin;D:/modelbox-win10-x64-1.5.3/workspace/yolov7_pet/bin/../dependence/lib;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3\envs\tensorflow;C:\Users\yanso\miniconda3\envs\tensorflow\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\usr\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Scripts;C:\Users\yanso\miniconda3\envs\tensorflow\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\Library\usr\bin;C:\Users\yanso\miniconda3\Library\bin;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin 

(tensorflow) D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet>modelbox.exe -c D:/modelbox-win10-x64-1.5.3/workspace/yolov7_pet/bin/../graph/modelbox.conf 
[2024-06-10 06:42:50,922][ WARN][    iva_config.cc:143 ] update vas url failed. Fault, no vas projectid or iva endpoint
open log file D:/modelbox-win10-x64-1.5.3/workspace/yolov7_pet/bin/../hilens_data_dir/log/modelbox.log failed, No error
input dims is:1,3,640,640,
output dims is:1,25200,7,

HTTP服务启动后可以在另一个终端进行请求测试，进入yolov7_pet工程目录data文件夹中使用test_http.py脚本发起HTTP请求进行测试：

(tensorflow) PS D:\modelbox-win10-x64-1.5.3> cd D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet\data                                                        
(tensorflow) PS D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet\data> python .\test_http.py                                                                 

================ Abyssinian_1.jpg ================
.\test_http.py:33: DeprecationWarning: tostring() is deprecated. Use tobytes() instead.
  img_str = cv2.imencode('.jpg', img_data)[1].tostring()
response:  {"det_result": "[[0.554308044910431, 0.1864600658416748, 0.7089953303337098, 0.3776256084442139, 0.82369065284729, 0.0]]"}

================ saint_bernard_143.jpg ================
response:  {"det_result": "[[0.46182055473327643, 0.30239262580871584, 0.8193012714385988, 0.4969032764434815, 0.7603430151939392, 1.0]]"}

四、小结

本章我们介绍了如何使用ModelBox开发一个动物目标检测的AI应用，我们只需要准备模型文件以及简单的配置即可创建一个HTTP服务。同时我们可以了解到图片标注、数据处理和模型训练方法，以及对应的推理应用逻辑。

点击关注，第一时间了解华为云新鲜技术~

微信关注我们

原文链接：https://my.oschina.net/u/4526289/blog/11209016

转载内容版权归作者及来源网站所有！

低调大师中文资讯倾力打造互联网数据资讯、行业资源、电子商务、移动互联网、网络营销平台。持续更新报道IT业界、互联网、市场资讯、驱动更新,是最及时权威的产业资讯及硬件资讯报道平台。

基于Ascend C的FlashAttention算子性能优化最佳实践

本文分享自华为云社区《基于Ascend C的FlashAttention算子性能优化最佳实践》，作者：昇腾CANN。 LLM的Attention部分处理给计算系统带来巨大的计算和访存压力。业界先后出现FlashAttention、FlashAttention2等算法，通过计算等价和切分有效降低HBM数据访问量。昇腾异构计算架构CANN针对昇腾AI处理器的片上内存和缓存大小，以及数据搬运通路，基于Ascend C算子编程语言优化实现FlashAttention融合算子，充分利用片上缓存，提升Attention处理性能。根据实测，在一些典型场景中CANN的FlashAttention算子相比小算子取得了5倍以上的性能提升，开发者可直接调用相关算子API接口使能大模型极致性能优化。本文针对FlashAttention反向融合算子的性能优化方案展开介绍，并通过优化实现了典型场景4倍左右的性能提升，希望对开发者优化此类基于Ascend C开发的融合算子带来启发。 FlashAttention算法简介在主流大模型网络模型中，大量使用典型的Multi-Head Attention结构，带来了巨...

2024-06-12

451

作者 |ttt 导读本文介绍了打点治理的概念和其对于数据质量保障的重要性，分享了日志中台在打点治理方面的解决方案与实践经验。文章从用户痛点和打点治理的难点出发，介绍了日志中台如何通过质量标准的制定、在线化流程的建设和相应的配套工具来解决这些问题。全文4808字，预计阅读时间13分钟。打点是指在网站或者APP中加入一些统计代码，通过日志记录用户在APP内触发的一系列行为，包括点击、滑动等。打点上报后汇聚成用户行为日志，用户行为日志可用于报表统计、AB Testing、个性化推荐等，是分析用户、调整策略、迭代产品的重要依据。打点治理是指在打点的生命周期内对其进行监控与管理，目标是确保数据的准确性、及时性、可比性、一致性、适用性和可获得性。在实践中我们经常会发现，日志数据总是会变得无效以至于无用、甚至有时新增打点也因为各种原因变得杂乱无章。因此从增量数据产生到存量数据维护的全流程，都是打点治理不可忽视的内容，也是日志中台关注的重点。日志中台建设并打磨打点平台，聚焦打点内容管理、管理打点全生命周期、推动打点规范设计、打点开发测试、打点效果验证等工具在业务侧落地。本文从打点平台角度，描...

2024-06-13

298

资源下载

更多资源

优质分享App

近一个月的开发和优化，本站点的第一个app全新上线。该app采用极致压缩，本体才4.36MB。系统里面做了大量数据访问、缓存优化。方便用户在手机上查看文章。后续会推出HarmonyOS的适配版本。

腾讯云软件源

为解决软件依赖安装时官方源访问速度慢的问题，腾讯云为一些软件搭建了缓存服务。您可以通过使用腾讯云软件源站来提升依赖包的安装速度。为了方便用户自由搭建服务架构，目前腾讯云软件源站支持公网访问和内网访问。

Spring

Spring框架（Spring Framework）是由Rod Johnson于2002年提出的开源Java企业级应用框架，旨在通过使用JavaBean替代传统EJB实现方式降低企业级编程开发的复杂性。该框架基于简单性、可测试性和松耦合性设计理念，提供核心容器、应用上下文、数据访问集成等模块，支持整合Hibernate、Struts等第三方框架，其适用范围不仅限于服务器端开发，绝大多数Java应用均可从中受益。

Rocky Linux

Rocky Linux（中文名：洛基）是由Gregory Kurtzer于2020年12月发起的企业级Linux发行版，作为CentOS稳定版停止维护后与RHEL（Red Hat Enterprise Linux）完全兼容的开源替代方案，由社区拥有并管理，支持x86_64、aarch64等架构。其通过重新编译RHEL源代码提供长期稳定性，采用模块化包装和SELinux安全架构，默认包含GNOME桌面环境及XFS文件系统，支持十年生命周期更新。