新手学习yolov8目标检测小记2--对比实验中经典模型库MMDetection使用方法（使用自己的数据集训练，并转换为yolo格式评价指标）

一、按照步骤环境配置

pip install timm==1.0.7 thop efficientnet_pytorch==0.7.1 einops grad-cam==1.4.8 dill==0.3.6 albumentations==1.4.11 pytorch_wavelets==1.3.0 tidecv PyWavelets -i https://pypi.tuna.tsinghua.edu.cn/simple
pip install -U openmim -i https://pypi.tuna.tsinghua.edu.cn/simple
mim install mmengine -i https://pypi.tuna.tsinghua.edu.cn/simple
mim install "mmcv==2.1.0" -i https://pypi.tuna.tsinghua.edu.cn/simple
pip install YOLO
pip install ultralytics
pip install -v -e.

二、自定义数据集放置

        我这里已经将数据集按照训练集、验证集、测试集=8:1:1划分好，具体的存放目录结构如下图所示。其中test2017、train2017、val2017存放图片，testlabels、trainlabels、vallabels存放标注文件txt，数据集格式转换后，在annotations文件中。

        YOLO格式转coco格式代码如下

import os
import cv2
import json
from tqdm import tqdm
from sklearn.model_selection import train_test_split
import argparse

# visdrone2019
classes = ['beibie1',
           'beibie2',
           'beibie3'
           ]

parser = argparse.ArgumentParser()
parser.add_argument('--image_path', default=r'E:\mmde\mmdetection-3.0.0\mmdetection-3.0.0\data\coco\val2017', type=str, help="path of images")
parser.add_argument('--label_path', default=r'E:\mmde\mmdetection-3.0.0\mmdetection-3.0.0\data\coco\vallabels', type=str, help="path of labels .txt")
parser.add_argument('--save_path', default='val.json', type=str,
                    help="if not split the dataset, give a path to a json file")
arg = parser.parse_args()


def yolo2coco(arg):
    print("Loading data from ", arg.image_path, arg.label_path)

    assert os.path.exists(arg.image_path)
    assert os.path.exists(arg.label_path)

    originImagesDir = arg.image_path
    originLabelsDir = arg.label_path
    # images dir name
    indexes = os.listdir(originImagesDir)

    dataset = {'categories': [], 'annotations': [], 'images': []}
    for i, cls in enumerate(classes, 0):
        dataset['categories'].append({'id': i, 'name': cls, 'supercategory': 'mark'})

    # 标注的id
    ann_id_cnt = 0
    for k, index in enumerate(tqdm(indexes)):
        # 支持 png jpg 格式的图片.
        txtFile = f'{index[:index.rfind(".")]}.txt'
        stem = index[:index.rfind(".")]
        # 读取图像的宽和高
        try:
            im = cv2.imread(os.path.join(originImagesDir, index))
            height, width, _ = im.shape
        except Exception as e:
            print(f'{os.path.join(originImagesDir, index)} read error.\nerror:{e}')
        # 添加图像的信息
        if not os.path.exists(os.path.join(originLabelsDir, txtFile)):
            # 如没标签，跳过，只保留图片信息.
            continue
        dataset['images'].append({'file_name': index,
                                  'id': stem,
                                  'width': width,
                                  'height': height})
        with open(os.path.join(originLabelsDir, txtFile), 'r') as fr:
            labelList = fr.readlines()
            for label in labelList:
                label = label.strip().split()
                x = float(label[1])
                y = float(label[2])
                w = float(label[3])
                h = float(label[4])

                # convert x,y,w,h to x1,y1,x2,y2
                H, W, _ = im.shape
                x1 = (x - w / 2) * W
                y1 = (y - h / 2) * H
                x2 = (x + w / 2) * W
                y2 = (y + h / 2) * H
                # 标签序号从0开始计算, coco2017数据集标号混乱，不管它了。
                cls_id = int(label[0])
                width = max(0, x2 - x1)
                height = max(0, y2 - y1)
                dataset['annotations'].append({
                    'area': width * height,
                    'bbox': [x1, y1, width, height],
                    'category_id': cls_id,
                    'id': ann_id_cnt,
                    'image_id': stem,
                    'iscrowd': 0,
                    # mask, 矩形是从左上角点按顺时针的四个顶点
                    'segmentation': [[x1, y1, x2, y1, x2, y2, x1, y2]]
                })
                ann_id_cnt += 1

    # 保存结果
    with open(arg.save_path, 'w') as f:
        json.dump(dataset, f)
        print('Save annotation to {}'.format(arg.save_path))


if __name__ == "__main__":
    yolo2coco(arg)

三、参数修改

        以faster-rcnn为例，查看文件configs/faster_rcnn/faster-rcnn_r50_fpn_1x_coco.py内容如下：

_base_ = [
    '../_base_/models/faster-rcnn_r50_fpn.py',
    '../_base_/datasets/coco_detection.py',
    '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
]

        根据显示内容，修改具体配置。（‘../_base_/default_runtime.py’无需修改）

        （1）到‘../_base_/models/faster-rcnn_r50_fpn.py’修改

num_classes为自己的实际数据集类别数。

        （2）到‘../_base_/datasets/coco_detection.py’，修改

dataset_type = 'CocoDataset'
data_root = 'data/coco/'

        由于我使用的是coco数据集格式，只需要改data_root为自己数据集的位置即可。并修改

scale为自己的图像尺寸大小，我的是scale=(640, 640)。接下来根据数据集修改ann_file和data_prefix，根据如上数据集的位置放置，我的修改后该文件完整内容如下：

# dataset settings
dataset_type = 'CocoDataset'
data_root = 'data/coco/'

# Example to use different file client
# Method 1: simply set the data root and let the file I/O module
# automatically infer from prefix (not support LMDB and Memcache yet)

# data_root = 's3://openmmlab/datasets/detection/coco/'

# Method 2: Use `backend_args`, `file_client_args` in versions before 3.0.0rc6
# backend_args = dict(
#     backend='petrel',
#     path_mapping=dict({
#         './data/': 's3://openmmlab/datasets/detection/',
#         'data/': 's3://openmmlab/datasets/detection/'
#     }))
backend_args = None

train_pipeline = [
    dict(type='LoadImageFromFile', backend_args=backend_args),
    dict(type='LoadAnnotations', with_bbox=True),
    dict(type='Resize', scale=(640, 640), keep_ratio=True),
    dict(type='RandomFlip', prob=0.5),
    dict(type='PackDetInputs')
]
test_pipeline = [
    dict(type='LoadImageFromFile', backend_args=backend_args),
    dict(type='Resize', scale=(640, 640), keep_ratio=True),
    # If you don't have a gt annotation, delete the pipeline
    dict(type='LoadAnnotations', with_bbox=True),
    dict(
        type='PackDetInputs',
        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
                   'scale_factor'))
]
train_dataloader = dict(
    batch_size=2,
    num_workers=2,
    persistent_workers=True,
    sampler=dict(type='DefaultSampler', shuffle=True),
    batch_sampler=dict(type='AspectRatioBatchSampler'),
    dataset=dict(
        type=dataset_type,
        data_root=data_root,
        ann_file='annotations/instances_train2017.json',
        data_prefix=dict(img='train2017/'),
        filter_cfg=dict(filter_empty_gt=True, min_size=32),
        pipeline=train_pipeline,
        backend_args=backend_args))
val_dataloader = dict(
    batch_size=1,
    num_workers=2,
    persistent_workers=True,
    drop_last=False,
    sampler=dict(type='DefaultSampler', shuffle=False),
    dataset=dict(
        type=dataset_type,
        data_root=data_root,
        ann_file='annotations/instances_val2017.json',
        data_prefix=dict(img='val2017/'),
        test_mode=True,
        pipeline=test_pipeline,
        backend_args=backend_args))
# test_dataloader = val_dataloader

val_evaluator = dict(
    type='CocoMetric',
    ann_file=data_root + 'annotations/instances_val2017.json',
    metric='bbox',
    format_only=False,
    backend_args=backend_args)
# test_evaluator = val_evaluator

# inference on test dataset and
# format the output results for submission.
test_dataloader = dict(
    batch_size=2,
    num_workers=2,
    persistent_workers=True,
    drop_last=False,
    sampler=dict(type='DefaultSampler', shuffle=False),
    dataset=dict(
        type=dataset_type,
        data_root=data_root,
        ann_file=data_root + 'annotations/instances_test2017.json',
        data_prefix=dict(img='test2017/'),
        test_mode=True,
        pipeline=test_pipeline))
test_evaluator = dict(
    type='CocoMetric',
    metric='bbox',
    format_only=True,
    ann_file=data_root + 'annotations/instances_test2017.json',
    outfile_prefix='./work_dirs/coco_detection/test')

        （3）到‘../_base_/schedules/schedule_1x.py’修改max_epochs为自己设置的最大训练轮次，其他的val_interval、lr、momentum、weight_decay，如果没有特别的要求可不修改。哦，根据自己的电脑情况，别忘了改base_batch_size。该文件整体内容如下：

# training schedule for 1x
train_cfg = dict(type='EpochBasedTrainLoop', max_epochs=100, val_interval=10)
val_cfg = dict(type='ValLoop')
test_cfg = dict(type='TestLoop')

# learning rate
param_scheduler = [
    dict(
        type='LinearLR', start_factor=0.001, by_epoch=False, begin=0, end=500),
    dict(
        type='MultiStepLR',
        begin=0,
        end=100,
        by_epoch=True,
        milestones=[67, 92],
        gamma=0.1)
]

# optimizer
optim_wrapper = dict(
    type='OptimWrapper',
    optimizer=dict(type='SGD', lr=0.01, momentum=0.937, weight_decay=0.0001))

# Default setting for scaling LR automatically
#   - `enable` means enable scaling LR automatically
#       or not by default.
#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
auto_scale_lr = dict(enable=False, base_batch_size=16)

四、命令训练

        使用如下命令训练：

python  tools/train.py configs/faster_rcnn/faster-rcnn_r50_fpn_1x_coco.py

       如果训练中断，使用resume继续（注意‘--work-dir work_dirs/faster-rcnn_r50_fpn_1x_coco’是训练结果输出的位置，epoch_21.pth是上次训练中断后，输出的最后一个pth，根据自己的实际情况修改）：

python tools/train.py configs/faster_rcnn/faster-rcnn_r50_fpn_1x_coco.py  --work-dir work_dirs/faster-rcnn_r50_fpn_1x_coco --resume work_dirs/faster-rcnn_r50_fpn_1x_coco/epoch_21.pth

五、转为YOLO格式的评价指标

        （1）找出最佳epoch

import os
import subprocess
import pickle
import numpy as np
import json
from prettytable import PrettyTable
from tqdm import tqdm

# 设置工作目录和模型文件路径
work_dir = "work_dirs/faster-rcnn_r50_fpn_1x_coco"
config_file = "configs/faster_rcnn/faster-rcnn_r50_fpn_1x_coco.py"  # 你的config文件路径

# 存放模型权重文件（epoch_1.pth 到 epoch_100.pth）
checkpoint_dir = os.path.join(work_dir, "")  # 假设检查点文件在 'checkpoints' 子文件夹下

# 遍历模型权重文件（epoch_1.pth, epoch_2.pth, ..., epoch_100.pth）
checkpoint_files = [f for f in os.listdir(checkpoint_dir) if f.endswith('.pth')]
checkpoint_files.sort(key=lambda x: int(x.split('_')[1].split('.')[0]))  # 按照 epoch 数字排序

# 用于存储评估结果
results = []

# 循环遍历每个模型文件
for checkpoint_file in tqdm(checkpoint_files, desc="Evaluating"):
    checkpoint_path = os.path.join(checkpoint_dir, checkpoint_file)

    # 设置输出 pkl 文件的路径
    output_pkl = f"res_{checkpoint_file.split('.')[0]}.pkl"

    # 运行 test.py 脚本进行模型评估
    command = [
        "python", "tools/test.py", config_file, checkpoint_path,
        "--out", output_pkl  # 输出 pkl 文件
    ]

    # 使用 subprocess 运行命令并捕获输出
    result = subprocess.run(command, capture_output=True, text=True)

    # 调试输出
    print(f"Evaluating {checkpoint_file}...")
    print(result.stdout)

    # 假设输出的评估结果包含了 mAP
    # 解析 mAP，假设它包含在 stdout 中，例如 "bbox_mAP: 0.45"
    for line in result.stdout.splitlines():
        if "bbox_mAP" in line:
            try:
                # 提取 mAP 分数
                bbox_mAP = float(line.split(":")[-1].strip())
                epoch = int(checkpoint_file.split("_")[1].split(".")[0])  # 获取 epoch 数字
                results.append((epoch, bbox_mAP, output_pkl))  # 存储结果 (epoch, mAP, pkl文件)
                break
            except ValueError:
                print(f"Error parsing bbox_mAP for {checkpoint_file}: {line}")
                continue

# 计算并输出最佳 epoch 和 mAP
if results:
    # 根据 mAP 找到最佳的 epoch
    best_epoch, best_mAP, best_pkl = max(results, key=lambda x: x[1])
    print(f"Best Epoch: {best_epoch}, Best bbox_mAP: {best_mAP:.4f}, Best Output pkl: {best_pkl}")
else:
    print("No valid results found. Please check the log outputs.")

# 计算所有 epoch 的 mAP 值
table = PrettyTable()
table.title = f"Evaluation Metrics"
table.field_names = ["Epoch", "bbox_mAP", "Output pkl"]

# 添加每个 epoch 的评估结果
for epoch, mAP, pkl_file in results:
    table.add_row([epoch, f"{mAP:.4f}", pkl_file])

print(table)

        （2）根据最佳epoch生成pkl文件 

python tools/test.py work_dirs/faster-rcnn_r50_fpn_1x_coco/faster-rcnn_r50_fpn_1x_coco.py work_dirs/faster-rcnn_r50_fpn_1x_coco/best_coco_bbox_mAP_epoch_90.pth --out res90.pkl

        （3）根据pkl文件输出对比参数，修改内容在‘def parse_opt():’，将内容改为实际的地址名称。

        完整代码如下：

import os, torch, cv2, math, tqdm, time, shutil, argparse, json, pickle
import numpy as np
from prettytable import PrettyTable


def clip_boxes(boxes, shape):
    # Clip boxes (xyxy) to image shape (height, width)
    if isinstance(boxes, torch.Tensor):  # faster individually
        boxes[..., 0].clamp_(0, shape[1])  # x1
        boxes[..., 1].clamp_(0, shape[0])  # y1
        boxes[..., 2].clamp_(0, shape[1])  # x2
        boxes[..., 3].clamp_(0, shape[0])  # y2
    else:  # np.array (faster grouped)
        boxes[..., [0, 2]] = boxes[..., [0, 2]].clip(0, shape[1])  # x1, x2
        boxes[..., [1, 3]] = boxes[..., [1, 3]].clip(0, shape[0])  # y1, y2


def scale_boxes(img1_shape, boxes, img0_shape, ratio_pad=None):
    # Rescale boxes (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]

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


def box_iou(box1, box2, eps=1e-7):
    """
    Calculate intersection-over-union (IoU) of boxes. Both sets of boxes are expected to be in (x1, y1, x2, y2) format.
    Based on https://github.com/pytorch/vision/blob/master/torchvision/ops/boxes.py
    Args:
        box1 (torch.Tensor): A tensor of shape (N, 4) representing N bounding boxes.
        box2 (torch.Tensor): A tensor of shape (M, 4) representing M bounding boxes.
        eps (float, optional): A small value to avoid division by zero. Defaults to 1e-7.
    Returns:
        (torch.Tensor): An NxM tensor containing the pairwise IoU values for every element in box1 and box2.
    """

    # NOTE: Need .float() to get accurate iou values
    # inter(N,M) = (rb(N,M,2) - lt(N,M,2)).clamp(0).prod(2)
    (a1, a2), (b1, b2) = box1.float().unsqueeze(1).chunk(2, 2), box2.float().unsqueeze(0).chunk(2, 2)
    inter = (torch.min(a2, b2) - torch.max(a1, b1)).clamp_(0).prod(2)

    # IoU = inter / (area1 + area2 - inter)
    return inter / ((a2 - a1).prod(2) + (b2 - b1).prod(2) - inter + eps)


def process_batch(detections, labels, iouv):
    """
    Return correct prediction matrix
    Arguments:
        detections (array[N, 6]), x1, y1, x2, y2, conf, class
        labels (array[M, 5]), class, x1, y1, x2, y2
    Returns:
        correct (array[N, 10]), for 10 IoU levels
    """
    correct = np.zeros((detections.shape[0], iouv.shape[0])).astype(bool)
    iou = box_iou(labels[:, 1:], detections[:, :4])
    correct_class = labels[:, 0:1] == detections[:, 5]
    for i in range(len(iouv)):
        x = torch.where((iou >= iouv[i]) & correct_class)  # IoU > threshold and classes match
        if x[0].shape[0]:
            matches = torch.cat((torch.stack(x, 1), iou[x[0], x[1]][:, None]), 1).cpu().numpy()  # [label, detect, iou]
            if x[0].shape[0] > 1:
                matches = matches[matches[:, 2].argsort()[::-1]]
                matches = matches[np.unique(matches[:, 1], return_index=True)[1]]
                # matches = matches[matches[:, 2].argsort()[::-1]]
                matches = matches[np.unique(matches[:, 0], return_index=True)[1]]
            correct[matches[:, 1].astype(int), i] = True
    return torch.tensor(correct, dtype=torch.bool, device=iouv.device)


def smooth(y, f=0.05):
    # Box filter of fraction f
    nf = round(len(y) * f * 2) // 2 + 1  # number of filter elements (must be odd)
    p = np.ones(nf // 2)  # ones padding
    yp = np.concatenate((p * y[0], y, p * y[-1]), 0)  # y padded
    return np.convolve(yp, np.ones(nf) / nf, mode='valid')  # y-smoothed


def ap_per_class(tp, conf, pred_cls, target_cls, plot=False, save_dir='.', names=(), eps=1e-16, prefix=''):
    """ Compute the average precision, given the recall and precision curves.
    Source: https://github.com/rafaelpadilla/Object-Detection-Metrics.
    # Arguments
        tp:  True positives (nparray, nx1 or nx10).
        conf:  Objectness value from 0-1 (nparray).
        pred_cls:  Predicted object classes (nparray).
        target_cls:  True object classes (nparray).
        plot:  Plot precision-recall curve at mAP@0.5
        save_dir:  Plot save directory
    # Returns
        The average precision as computed in py-faster-rcnn.
    """

    # Sort by objectness
    i = np.argsort(-conf)
    tp, conf, pred_cls = tp[i], conf[i], pred_cls[i]

    # Find unique classes
    unique_classes, nt = np.unique(target_cls, return_counts=True)
    nc = unique_classes.shape[0]  # number of classes, number of detections

    # Create Precision-Recall curve and compute AP for each class
    px, py = np.linspace(0, 1, 1000), []  # for plotting
    ap, p, r = np.zeros((nc, tp.shape[1])), np.zeros((nc, 1000)), np.zeros((nc, 1000))
    for ci, c in enumerate(unique_classes):
        i = pred_cls == c
        n_l = nt[ci]  # number of labels
        n_p = i.sum()  # number of predictions
        if n_p == 0 or n_l == 0:
            continue

        # Accumulate FPs and TPs
        fpc = (1 - tp[i]).cumsum(0)
        tpc = tp[i].cumsum(0)

        # Recall
        recall = tpc / (n_l + eps)  # recall curve
        r[ci] = np.interp(-px, -conf[i], recall[:, 0], left=0)  # negative x, xp because xp decreases

        # Precision
        precision = tpc / (tpc + fpc)  # precision curve
        p[ci] = np.interp(-px, -conf[i], precision[:, 0], left=1)  # p at pr_score

        # AP from recall-precision curve
        for j in range(tp.shape[1]):
            ap[ci, j], mpre, mrec = compute_ap(recall[:, j], precision[:, j])
            if plot and j == 0:
                py.append(np.interp(px, mrec, mpre))  # precision at mAP@0.5

    # Compute F1 (harmonic mean of precision and recall)
    f1 = 2 * p * r / (p + r + eps)

    i = smooth(f1.mean(0), 0.1).argmax()  # max F1 index
    p, r, f1 = p[:, i], r[:, i], f1[:, i]
    tp = (r * nt).round()  # true positives
    fp = (tp / (p + eps) - tp).round()  # false positives
    return tp, fp, p, r, f1, ap, unique_classes.astype(int)


def compute_ap(recall, precision):
    """ Compute the average precision, given the recall and precision curves
    # Arguments
        recall:    The recall curve (list)
        precision: The precision curve (list)
    # Returns
        Average precision, precision curve, recall curve
    """

    # Append sentinel values to beginning and end
    mrec = np.concatenate(([0.0], recall, [1.0]))
    mpre = np.concatenate(([1.0], precision, [0.0]))

    # Compute the precision envelope
    mpre = np.flip(np.maximum.accumulate(np.flip(mpre)))

    # Integrate area under curve
    method = 'interp'  # methods: 'continuous', 'interp'
    if method == 'interp':
        x = np.linspace(0, 1, 101)  # 101-point interp (COCO)
        ap = np.trapz(np.interp(x, mrec, mpre), x)  # integrate
    else:  # 'continuous'
        i = np.where(mrec[1:] != mrec[:-1])[0]  # points where x axis (recall) changes
        ap = np.sum((mrec[i + 1] - mrec[i]) * mpre[i + 1])  # area under curve

    return ap, mpre, mrec


def parse_opt():
    parser = argparse.ArgumentParser()
    parser.add_argument('--label_coco', type=str, default='E:/mmde/mmdetection-3.0.0/mmdetection-3.0.0/test.json',
                        help='label coco path')
    # parser.add_argument('--pred_coco', type=str, default='runs/val/exp/predictions.json', help='pred coco path')
    parser.add_argument('--pred_coco', type=str, default='E:/mmde/mmdetection-3.0.0/mmdetection-3.0.0/res90.pkl', help='pred coco path')
    parser.add_argument('--iou', type=float, default=0.7, help='iou threshold')
    parser.add_argument('--conf', type=float, default=0.001, help='conf threshold')
    opt = parser.parse_known_args()[0]
    return opt


if __name__ == '__main__':
    opt = parse_opt()

    iouv = torch.linspace(0.5, 0.95, 10)  # iou vector for mAP@0.5:0.95
    niou = iouv.numel()
    stats = []

    label_coco_json_path, pred_coco_json_path = opt.label_coco, opt.pred_coco
    with open(label_coco_json_path) as f:
        label = json.load(f)

    classes = []
    for data in label['categories']:
        classes.append(data['name'])

    image_id_hw_dict = {}
    for data in label['images']:
        image_id_hw_dict[data['id']] = [data['height'], data['width']]

    label_id_dict = {}
    for data in tqdm.tqdm(label['annotations'], desc='Process label...'):
        if data['image_id'] not in label_id_dict:
            label_id_dict[data['image_id']] = []

        category_id = data['category_id']
        x_min, y_min, w, h = data['bbox'][0], data['bbox'][1], data['bbox'][2], data['bbox'][3]
        x_max, y_max = x_min + w, y_min + h
        label_id_dict[data['image_id']].append(np.array([int(category_id), x_min, y_min, x_max, y_max]))

    if pred_coco_json_path.endswith('json'):
        with open(pred_coco_json_path) as f:
            pred = json.load(f)
        pred_id_dict = {}
        for data in tqdm.tqdm(pred, desc='Process pred...'):
            if data['image_id'] not in pred_id_dict:
                pred_id_dict[data['image_id']] = []

            score = data['score']
            category_id = data['category_id']
            x_min, y_min, w, h = data['bbox'][0], data['bbox'][1], data['bbox'][2], data['bbox'][3]
            x_max, y_max = x_min + w, y_min + h

            pred_id_dict[data['image_id']].append(
                np.array([x_min, y_min, x_max, y_max, float(score), int(category_id)]))
    else:
        with open(pred_coco_json_path, 'rb') as f:
            pred = pickle.load(f)
        pred_id_dict = {}
        for data in tqdm.tqdm(pred, desc='Process pred...'):
            image_id = os.path.splitext(os.path.basename(data['img_path']))[0]
            if image_id not in pred_id_dict:
                pred_id_dict[image_id] = []

            for i in range(data['pred_instances']['labels'].size(0)):
                score = data['pred_instances']['scores'][i]
                category_id = data['pred_instances']['labels'][i]
                bboxes = data['pred_instances']['bboxes'][i]

                x_min, y_min, x_max, y_max = bboxes.cpu().detach().numpy()
                # x_min, x_max = x_min / data['scale_factor'][0], x_max / data['scale_factor'][0]
                # y_min, y_max = y_min / data['scale_factor'][1], y_max / data['scale_factor'][1]

                pred_id_dict[image_id].append(np.array([x_min, y_min, x_max, y_max, float(score), int(category_id)]))

    for idx, image_id in enumerate(tqdm.tqdm(list(image_id_hw_dict.keys()), desc="Cal mAP...")):
        label = np.array(label_id_dict[image_id])

        if image_id not in pred_id_dict:
            pred = np.empty((0, 6))
        else:
            pred = torch.from_numpy(np.array(pred_id_dict[image_id]))

        nl, npr = label.shape[0], pred.shape[0]
        correct = torch.zeros(npr, niou, dtype=torch.bool)
        if npr == 0:
            if nl:
                stats.append((correct, *torch.zeros((2, 0)), torch.from_numpy(label[:, 0])))
            continue

        if nl:
            correct = process_batch(pred, torch.from_numpy(label), iouv)
        stats.append((correct, pred[:, 4], pred[:, 5], torch.from_numpy(label[:, 0])))

    stats = [torch.cat(x, 0).cpu().numpy() for x in zip(*stats)]
    tp, fp, p, r, f1, ap, ap_class = ap_per_class(*stats)
    print(f'precision:{p}')
    print(f'recall:{r}')
    print(f'mAP@0.5:{ap[:, 0]}')

    table = PrettyTable()
    table.title = f"Metrice"
    table.field_names = ["Classes", 'Precision', 'Recall', 'mAP50', 'mAP50-95']
    table.add_row(['all', f'{np.mean(p):.3f}', f'{np.mean(r):.3f}', f'{np.mean(ap[:, 0]):.3f}', f'{np.mean(ap):.3f}'])
    for cls_idx, classes in enumerate(classes):
        table.add_row([classes, f'{p[cls_idx]:.3f}', f'{r[cls_idx]:.3f}', f'{ap[cls_idx, 0]:.3f}',
                       f'{ap[cls_idx, :].mean():.3f}'])
    print(table)

六、输出结果