您现在的位置是：首页 >学无止境 >Rk1126 实现 yolov5 6.2 推理网站首页学无止境

Rk1126 实现 yolov5 6.2 推理

J .. 2024-07-01 11:59:42
简介Rk1126 实现 yolov5 6.2 推理

在这里插入图片描述

基于 RK1126 实现 yolov5 6.2 推理.

转换 ONNX

  1. python export.py --weights ./weights/yolov5s.pt --img 640 --batch 1 --include onnx --simplify

安装 rk 环境

  1. 安装部分参考网上, 有很多. 参考: https://github.com/rockchip-linux/rknpu

转换 RK模型 并验证

  1. yolov562_to_rknn_3_4.py ( s/m/l/x ..，输出节点不同，使用 netror 查看 )

    # -*- coding: utf-8 -*-
"""
Created on Wed Oct 12 18:24:38 2022

@author: bobod
"""


import os
import numpy as np
import cv2
from rknn.api import RKNN


ONNX_MODEL = './weights/yolov5s_v6.2.onnx'
RKNN_MODEL = './weights/yolov5s_v6.2.rknn'
IMG_PATH = './000000102411.jpg'
DATASET = './dataset.txt'

QUANTIZE_ON = True

BOX_THRESH = 0.5
NMS_THRESH = 0.6
IMG_SIZE = (640, 640) # (width, height), such as (1280, 736)

SHAPES =((0.0, 0.0), (0.0, 0.0)) #1 scale_coords
SHAPE =(0,0)

CLASSES = ("person", "bicycle", "car","motorbike ","aeroplane ","bus ","train","truck ","boat","traffic light",
           "fire hydrant","stop sign ","parking meter","bench","bird","cat","dog ","horse ","sheep","cow","elephant",
           "bear","zebra ","giraffe","backpack","umbrella","handbag","tie","suitcase","frisbee","skis","snowboard","sports ball","kite",
           "baseball bat","baseball glove","skateboard","surfboard","tennis racket","bottle","wine glass","cup","fork","knife ",
           "spoon","bowl","banana","apple","sandwich","orange","broccoli","carrot","hot dog","pizza ","donut","cake","chair","sofa",
           "pottedplant","bed","diningtable","toilet ","tvmonitor","laptop	","mouse	","remote ","keyboard ","cell phone","microwave ",
           "oven ","toaster","sink","refrigerator ","book","clock","vase","scissors ","teddy bear ","hair drier", "toothbrush ")

def sigmoid(x):
    return 1 / (1 + np.exp(-x))

def xywh2xyxy(x):
    # Convert [x, y, w, h] to [x1, y1, x2, y2]
    y = np.copy(x)
    y[:, 0] = x[:, 0] - x[:, 2] / 2  # top left x
    y[:, 1] = x[:, 1] - x[:, 3] / 2  # top left y
    y[:, 2] = x[:, 0] + x[:, 2] / 2  # bottom right x
    y[:, 3] = x[:, 1] + x[:, 3] / 2  # bottom right y
    return y

def process(input, mask, anchors):

    anchors = [anchors[i] for i in mask]
    grid_h, grid_w = map(int, input.shape[0:2])

    box_confidence = sigmoid(input[..., 4])
    box_confidence = np.expand_dims(box_confidence, axis=-1)

    box_class_probs = sigmoid(input[..., 5:])

    box_xy = sigmoid(input[..., :2])*2 - 0.5

    col = np.tile(np.arange(0, grid_w), grid_h).reshape(-1, grid_w)
    row = np.tile(np.arange(0, grid_h).reshape(-1, 1), grid_w)
    col = col.reshape(grid_h, grid_w, 1, 1).repeat(3, axis=-2)
    row = row.reshape(grid_h, grid_w, 1, 1).repeat(3, axis=-2)
    grid = np.concatenate((col, row), axis=-1)
    box_xy += grid
    box_xy *= (int(IMG_SIZE[1]/grid_h), int(IMG_SIZE[0]/grid_w))

    box_wh = pow(sigmoid(input[..., 2:4])*2, 2)
    box_wh = box_wh * anchors

    box = np.concatenate((box_xy, box_wh), axis=-1)

    return box, box_confidence, box_class_probs

def filter_boxes(boxes, box_confidences, box_class_probs):
    """Filter boxes with box threshold. It's a bit different with origin yolov5 post process!
    # Arguments
        boxes: ndarray, boxes of objects.
        box_confidences: ndarray, confidences of objects.
        box_class_probs: ndarray, class_probs of objects.
    # Returns
        boxes: ndarray, filtered boxes.
        classes: ndarray, classes for boxes.
        scores: ndarray, scores for boxes.
    """
    boxes = boxes.reshape(-1, 4)
    box_confidences = box_confidences.reshape(-1)
    box_class_probs = box_class_probs.reshape(-1, box_class_probs.shape[-1])

    _box_pos = np.where(box_confidences >= BOX_THRESH)
    boxes = boxes[_box_pos]
    box_confidences = box_confidences[_box_pos]
    box_class_probs = box_class_probs[_box_pos]

    class_max_score = np.max(box_class_probs, axis=-1)
    classes = np.argmax(box_class_probs, axis=-1)
    _class_pos = np.where(class_max_score* box_confidences >= BOX_THRESH)

    boxes = boxes[_class_pos]
    classes = classes[_class_pos]
    scores = (class_max_score* box_confidences)[_class_pos]

    return boxes, classes, scores

def nms_boxes(boxes, scores):
    """Suppress non-maximal boxes.
    # Arguments
        boxes: ndarray, boxes of objects.
        scores: ndarray, scores of objects.
    # Returns
        keep: ndarray, index of effective boxes.
    """
    x = boxes[:, 0]
    y = boxes[:, 1]
    w = boxes[:, 2] - boxes[:, 0]
    h = boxes[:, 3] - boxes[:, 1]

    areas = w * h
    order = scores.argsort()[::-1]

    keep = []
    while order.size > 0:
        i = order[0]
        keep.append(i)

        xx1 = np.maximum(x[i], x[order[1:]])
        yy1 = np.maximum(y[i], y[order[1:]])
        xx2 = np.minimum(x[i] + w[i], x[order[1:]] + w[order[1:]])
        yy2 = np.minimum(y[i] + h[i], y[order[1:]] + h[order[1:]])

        w1 = np.maximum(0.0, xx2 - xx1 + 0.00001)
        h1 = np.maximum(0.0, yy2 - yy1 + 0.00001)
        inter = w1 * h1

        ovr = inter / (areas[i] + areas[order[1:]] - inter)
        inds = np.where(ovr <= NMS_THRESH)[0]
        order = order[inds + 1]
    keep = np.array(keep)
    return keep


def yolov5_post_process(input_data):
    masks = [[0, 1, 2], [3, 4, 5], [6, 7, 8]]
    anchors = [[10, 13], [16, 30], [33, 23], [30, 61], [62, 45],
              [59, 119], [116, 90], [156, 198], [373, 326]]

    boxes, classes, scores = [], [], []
    for input,mask in zip(input_data, masks):
        b, c, s = process(input, mask, anchors)
        b, c, s = filter_boxes(b, c, s)
        boxes.append(b)
        classes.append(c)
        scores.append(s)

    boxes = np.concatenate(boxes)
    boxes = xywh2xyxy(boxes)
    classes = np.concatenate(classes)
    scores = np.concatenate(scores)

    nboxes, nclasses, nscores = [], [], []
    for c in set(classes):
        inds = np.where(classes == c)
        b = boxes[inds]
        c = classes[inds]
        s = scores[inds]

        keep = nms_boxes(b, s)

        nboxes.append(b[keep])
        nclasses.append(c[keep])
        nscores.append(s[keep])

    if not nclasses and not nscores:
        return None, None, None

    boxes = np.concatenate(nboxes)
    scale_coords(IMG_SIZE, boxes, SHAPE, SHAPES) #2
    classes = np.concatenate(nclasses)
    scores = np.concatenate(nscores)

    return boxes, classes, scores

def draw(image, boxes, scores, classes):
    """Draw the boxes on the image.
    # Argument:
        image: original image.
        boxes: ndarray, boxes of objects.
        classes: ndarray, classes of objects.
        scores: ndarray, scores of objects.
        all_classes: all classes name.
    """
    for box, score, cl in zip(boxes, scores, classes):
        left, top, right, bottom = box
        print('class: {}, score: {}'.format(CLASSES[cl], score))
        print('box coordinate left,top,right,bottom: [{}, {}, {}, {}]'.format(left, top, right, bottom))
        left = int(left)
        top = int(top)
        right = int(right)
        bottom = int(bottom)

        cv2.rectangle(image, (left, top), (right, bottom), (255, 0, 0), 2)
        cv2.putText(image, '{0} {1:.2f}'.format(CLASSES[cl], score),
                    (left, top - 6),
                    cv2.FONT_HERSHEY_SIMPLEX,
                    0.6, (0, 0, 255), 2)


def letterbox(im, new_shape=(640, 640), color=(0, 0, 0)):
    # Resize and pad image while meeting stride-multiple constraints
    shape = im.shape[:2]  # current shape [height, width]
    if isinstance(new_shape, int):
        new_shape = (new_shape, new_shape)

    # Scale ratio (new / old)
    r = min(new_shape[0] / shape[0], new_shape[1] / shape[1])

    # Compute padding
    ratio = r, r  # width, height ratios
    new_unpad = int(round(shape[1] * r)), int(round(shape[0] * r))
    dw, dh = new_shape[1] - new_unpad[0], new_shape[0] - new_unpad[1]  # wh padding

    dw /= 2  # divide padding into 2 sides
    dh /= 2

    if shape[::-1] != new_unpad:  # resize
        im = cv2.resize(im, new_unpad, interpolation=cv2.INTER_LINEAR)
    top, bottom = int(round(dh - 0.1)), int(round(dh + 0.1))
    left, right = int(round(dw - 0.1)), int(round(dw + 0.1))
    im = cv2.copyMakeBorder(im, top, bottom, left, right, cv2.BORDER_CONSTANT, value=color)  # add border
    return im, ratio, (dw, dh)

#3
def scale_coords(img1_shape, coords, img0_shape, ratio_pad=None):
    # Rescale coords (xyxy) from img1_shape to img0_shape
    if ratio_pad is None:  # calculate from img0_shape
        gain = min(img1_shape[0] / img0_shape[0], img1_shape[1] / img0_shape[1])  # gain  = old / new
        pad = (img1_shape[1] - img0_shape[1] * gain) / 2, (img1_shape[0] - img0_shape[0] * gain) / 2  # wh padding
    else:
        gain = ratio_pad[0][0]
        pad = ratio_pad[1]

    coords[:, [0, 2]] -= pad[0]  # x padding
    coords[:, [1, 3]] -= pad[1]  # y padding
    coords[:, :4] /= gain
    clip_coords(coords, img0_shape)
    return coords


def clip_coords(boxes, shape):
    # Clip bounding xyxy bounding boxes to image shape (height, width)
    boxes[:, [0, 2]] = boxes[:, [0, 2]].clip(0, shape[1])  # x1, x2
    boxes[:, [1, 3]] = boxes[:, [1, 3]].clip(0, shape[0])  # y1, y2
    
if __name__ == '__main__':

    # Create RKNN object
    rknn = RKNN(verbose=False)

    if not os.path.exists(ONNX_MODEL):
        print('model not exist')
        exit(-1)

    _force_builtin_perm = False
    # pre-process config
    print('--> Config model')
    rknn.config(
                reorder_channel='0 1 2',
                mean_values=[[0, 0, 0]],
                std_values=[[255, 255, 255]],
                optimization_level=3,
                #target_platform = 'rk1808',
                # target_platform='rv1109',
                target_platform = 'rv1126',
                quantize_input_node= QUANTIZE_ON,
                output_optimize=1,
                force_builtin_perm=_force_builtin_perm)
    print('done')

    # Load ONNX model
    print('--> Loading model')
    #ret = rknn.load_pytorch(model=PT_MODEL, input_size_list=[[3,IMG_SIZE[1], IMG_SIZE[0]]])
    ret = rknn.load_onnx(model=ONNX_MODEL, outputs=['output', '391', '402'])
    if ret != 0:
        print('Load yolov5 failed!')
        exit(ret)
    print('done')

    # Build model
    print('--> Building model')
    ret = rknn.build(do_quantization=QUANTIZE_ON, dataset=DATASET, pre_compile=False)
    if ret != 0:
        print('Build yolov5 failed!')
        exit(ret)
    print('done')

    # Export RKNN model
    print('--> Export RKNN model')
    ret = rknn.export_rknn(RKNN_MODEL)
    if ret != 0:
        print('Export yolov5rknn failed!')
        exit(ret)
    print('done')

    # init runtime environment
    print('--> Init runtime environment')
    ret = rknn.init_runtime() 
    #ret = rknn.init_runtime('rv1126', device_id='bab4d7a824f04867')
    # ret = rknn.init_runtime('rv1109', device_id='1109')
    # ret = rknn.init_runtime('rk1808', device_id='1808')
    if ret != 0:
        print('Init runtime environment failed')
        exit(ret)
    print('done')

    # Set inputs
    original_img = cv2.imread(IMG_PATH)#4
    img, ratio, pad = letterbox(original_img, new_shape=(IMG_SIZE[1], IMG_SIZE[0]))
    SHAPES=(ratio,pad)
    SHAPE=(original_img.shape[0],original_img.shape[1])
    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

    # Inference
    print('--> Running model')
    outputs = rknn.inference(inputs=[img], inputs_pass_through=[0 if not _force_builtin_perm else 1])

    # post process
    input0_data = outputs[0]
    input1_data = outputs[1]
    input2_data = outputs[2]

    input0_data = input0_data.reshape([3,-1]+list(input0_data.shape[-2:]))
    input1_data = input1_data.reshape([3,-1]+list(input1_data.shape[-2:]))
    input2_data = input2_data.reshape([3,-1]+list(input2_data.shape[-2:]))

    input_data = list()
    input_data.append(np.transpose(input0_data, (2, 3, 0, 1)))
    input_data.append(np.transpose(input1_data, (2, 3, 0, 1)))
    input_data.append(np.transpose(input2_data, (2, 3, 0, 1)))


    boxes, classes, scores = yolov5_post_process(input_data)


    if boxes is not None:
        draw(original_img, boxes, scores, classes)
    cv2.imwrite("result.jpg", original_img)

在这里插入图片描述

混合量化

  1. 首先做精度分析, 量化 build 之后， 调用精度分析函数。后来尝试在 hybrid_quantization_step2 后调用，结果总是不成功 , 原因未知, 有知道的大佬望告知一下.
  2. 精度分析: entire_qnt ( 完全量化结果 ) ， fp32 ( fp32结果 ) ，individual_qnt ( 逐层量化结果,即输入为float, 排除累计误差 ), entire_qnt_error_analysis.txt individual_qnt_error_analysis.txt ( 完全量化和逐层量化分析结果 (欧式距离和余弦距离) )
  3. 这里需要注意, DATASET 只能有一行数据.
    ...
    # Build model
    print('--> Building model')
    ret = rknn.build(do_quantization=QUANTIZE_ON, dataset=DATASET, pre_compile=False)
    if ret != 0:
        print('Build yolov5 failed!')
        exit(ret)
    print('done')

    print('--> Accuracy analysis')
    ret = rknn.accuracy_analysis(inputs=DATASET1,output_dir="./output_dir")
    if ret != 0:
        print('accuracy_analysis failed!')
        exit(ret)
    print('done')
  4. 生成混合量化配置文件, 调用 rknn.hybrid_quantization_step1 会得到 torchjitexport.data ， torchjitexport.json， torchjitexport.quantization.cfg 3个文件.
  5. 根据精度分析结果，torchjitexport.quantization.cfg 将不想量化的层添加到自定义层中.
    # add layer name and corresponding quantized_dtype to customized_quantize_layers, e.g conv2_3: float32
customized_quantize_layers: {
    "Conv_Conv_0_187":float32,
    "Sigmoid_Sigmoid_1_188_Mul_Mul_2_172":float32,
    "Conv_Conv_3_171":float32,
    ...
}
...
  6. 调用 rknn.hybrid_quantization_step2 导出rknn模型
    ...
ret = rknn.hybrid_quantization_step2(model_input='./torchjitexport.json',data_input='./torchjitexport.data', model_quantization_cfg='./torchjitexport.quantization.cfg',dataset=DATASET, pre_compile=False)
    if ret != 0:
        print("hybrid_quantization_step2 failed. ")
        exit(ret)

# Export RKNN model
print('--> Export RKNN model')
ret = rknn.export_rknn(RKNN_MODEL)
if ret != 0:
    print('Export yolov5rknn failed!')
    exit(ret)
print('done')
  7. 重复测试模型，找出一个平衡速度和准确率的模型。
  8. 验证没问题了，就可以连接板子尝试啦，可以将 pre_compile=True, 提高初始化速度. 不过不能在仿真环境中测试.

参考

  1. https://github.com/shaoshengsong/rockchip_rknn_yolov5
  2. https://github.com/rockchip-linux/rknpu
风语者！平时喜欢研究各种技术，目前在从事后端开发工作，热爱生活、热爱工作。

站长推荐