yolov7-face关于widerface-val数据集的评测

1. 我复现了yolov7s-face的精度，我没有加载预训练模型，重新训的，300个epoch,最优的模型精度比官方的点还要高一点，下面是我模型的精度和官方的精度:
   我训出来模型的指标:
   yolov7s 训练300epoch best.pt
   ==================== Results ====================
   Easy Val AP: 0.9510914378803449
   Medium Val AP: 0.9349033662833984
   Hard Val AP: 0.8550381774914542
   =================================================
   官方模型的指标:
   ==================== Results ====================
   Easy Val AP: 0.9481596778871507
   Medium Val AP: 0.9314085577436426
   Hard Val AP: 0.8516288529133722
   =================================================
   2.如何对val数据集进行评测，先改一下detect.py代码生成用于评测精度的txt文件，代码如下:

import argparse
import time
from pathlib import Path

import os
import copy
import cv2
import torch
import torch.backends.cudnn as cudnn
from numpy import random

from models.experimental import attempt_load
from utils.datasets import LoadStreams, LoadImages
from utils.general import check_img_size, check_requirements, check_imshow, non_max_suppression, apply_classifier, \
    scale_coords, xyxy2xywh, strip_optimizer, set_logging, increment_path, save_one_box
from utils.plots import colors, plot_one_box
from utils.torch_utils import select_device, load_classifier, time_synchronized


def detect(opt):
    source, weights, view_img, save_txt, imgsz, save_txt_tidl, kpt_label ,widerface,save_widerface= opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size, opt.save_txt_tidl, opt.kpt_label,opt.widerface,opt.save_widerface
    save_img = not opt.nosave and not source.endswith('.txt')  # save inference images
    webcam = source.isnumeric() or source.endswith('.txt') or source.lower().startswith(
        ('rtsp://', 'rtmp://', 'http://', 'https://'))

    # Directories
    save_dir = increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok)  # increment run
    (save_dir / 'labels' if (save_txt or save_txt_tidl) else save_dir).mkdir(parents=True, exist_ok=True)  # make dir

    # Initialize
    set_logging()
    device = select_device(opt.device)
    half = device.type != 'cpu' and not save_txt_tidl  # half precision only supported on CUDA

    # Load model
    model = attempt_load(weights, map_location=device)  # load FP32 model
    stride = int(model.stride.max())  # model stride
    if isinstance(imgsz, (list,tuple)):
        assert len(imgsz) ==2; "height and width of image has to be specified"
        imgsz[0] = check_img_size(imgsz[0], s=stride)
        imgsz[1] = check_img_size(imgsz[1], s=stride)
    else:
        imgsz = check_img_size(imgsz, s=stride)  # check img_size
    names = model.module.names if hasattr(model, 'module') else model.names  # get class names
    if half:
        model.half()  # to FP16

    # Second-stage classifier
    classify = False
    if classify:
        modelc = load_classifier(name='resnet101', n=2)  # initialize
        modelc.load_state_dict(torch.load('weights/resnet101.pt', map_location=device)['model']).to(device).eval()

    # Set Dataloader
    vid_path, vid_writer = None, None
    if webcam:
        view_img = check_imshow()
        cudnn.benchmark = True  # set True to speed up constant image size inference
        dataset = LoadStreams(source, img_size=imgsz, stride=stride)
    else:
        dataset = LoadImages(source, img_size=imgsz, stride=stride)

    # Run inference
    if device.type != 'cpu':
        model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.parameters())))  # run once
    t0 = time.time()
    for path, img, im0s, vid_cap in dataset:
        img = torch.from_numpy(img).to(device)
        img = img.half() if half else img.float()  # uint8 to fp16/32
        img /= 255.0  # 0 - 255 to 0.0 - 1.0
        if img.ndimension() == 3:
            img = img.unsqueeze(0)

        # Inference
        t1 = time_synchronized()
        pred = model(img, augment=opt.augment)[0]
        print(pred[...,4].max())
        # Apply NMS
        pred = non_max_suppression(pred, opt.conf_thres, opt.iou_thres, classes=opt.classes, agnostic=opt.agnostic_nms, kpt_label=kpt_label)
        t2 = time_synchronized()

        # Apply Classifier
        if classify:
            pred = apply_classifier(pred, modelc, img, im0s)

        # Process detections
        for i, det in enumerate(pred):  # detections per image
            if webcam:  # batch_size >= 1
                p, s, im0, frame = path[i], '%g: ' % i, im0s[i].copy(), dataset.count
            else:
                p, s, im0, frame = path, '', im0s.copy(), getattr(dataset, 'frame', 0)

            p = Path(p)  # to Path
            save_path = str(save_dir / p.name)  # img.jpg
            txt_path = str(save_dir / 'labels' / p.stem) + ('' if dataset.mode == 'image' else f'_{frame}')  # img.txt
            s += '%gx%g ' % img.shape[2:]  # print string
            gn = torch.tensor(im0.shape)[[1, 0, 1, 0]]  # normalization gain whwh
            # print("det----",det)
            # print("det.Size()---",det.shape[0])
            # exit(-1)
            if det.shape[0] == 0:
                # print("det is None:====",det)
                # exit(-1)
                os.makedirs(save_widerface,exist_ok=True)
                save_widerface_txt = str(Path(save_widerface)/Path(Path(p).name[:-3]+"txt"))
                with open(save_widerface_txt,"w") as fwider:
                    widerface_file_name = Path(p).name[:-4] + "\n"
                    print("=========",widerface_file_name)
                    # exit(-1)
                    fwider.write(widerface_file_name)
                    fwider.write(str(0)+"\n")
            if len(det):
                # Rescale boxes from img_size to im0 size
                scale_coords(img.shape[2:], det[:, :4], im0.shape, kpt_label=False)
                scale_coords(img.shape[2:], det[:, 6:], im0.shape, kpt_label=kpt_label, step=3)

                # Print results
                for c in det[:, 5].unique():
                    n = (det[:, 5] == c).sum()  # detections per class
                    s += f"{n} {names[int(c)]}{'s' * (n > 1)}, "  # add to string
                    

                if widerface:
                    os.makedirs(save_widerface,exist_ok=True)
                    save_widerface_txt = str(Path(save_widerface)/Path(Path(p).name[:-3]+"txt"))
                    widerface_file_name = Path(p).name[:-4] + "\n"
                    
                    widerface_bboxs_num = str(len(pred[0])) +"\n"
                            
                    with open(save_widerface_txt,"a") as fwider:
                        fwider.write(widerface_file_name)
                        fwider.write(widerface_bboxs_num)
    
                # Write results
                for det_index, (*xyxy, conf, cls) in enumerate((det[:,:6])):
                    if widerface:
                        if cls == 0:
                            os.makedirs(save_widerface,exist_ok=True)
                            save_widerface_txt = str(Path(save_widerface)/Path(Path(p).name[:-3]+"txt"))
                            #widerface_file_name = Path(p).name + "\n"
                            #widerface_bboxs_num = str(len(pred[0])) +"\n"
                            x1 = int(xyxy[0]+0.5)
                            y1 = int(xyxy[1]+0.5)
                            x2 = int(xyxy[2]+0.5)
                            y2 = int(xyxy[3]+0.5)
                            with open(save_widerface_txt,"a") as fwider:
                                #fwider.write(widerface_file_name)
                                #fwider.write(widerface_bboxs_num)
                                fwider.write("%d %d %d %d %.03f"%(x1,y1,x2-x1,y2-y1,conf if conf<=1 else 1)+"\n")
                    if save_txt:  # Write to file
                        xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist()  # normalized xywh
                        line = (cls, *xywh, conf) if opt.save_conf else (cls, *xywh)  # label format
                        with open(txt_path + '.txt', 'a') as f:
                            f.write(('%g ' * len(line)).rstrip() % line + '\n')

                    # if save_img or opt.save_crop or view_img:  # Add bbox to image
                    #     c = int(cls)  # integer class
                    #     label = None if opt.hide_labels else (names[c] if opt.hide_conf else f'{names[c]} {conf:.2f}')
                    #     kpts = det[det_index, 6:]
                    #     plot_one_box(xyxy, im0, label=label, color=colors(c, True), line_thickness=opt.line_thickness, kpt_label=kpt_label, kpts=kpts, steps=3, orig_shape=im0.shape[:2])
                    #     if opt.save_crop:
                    #         save_one_box(xyxy, im0s, file=save_dir / 'crops' / names[c] / f'{p.stem}.jpg', BGR=True)


                if save_txt_tidl:  # Write to file in tidl dump format
                    for *xyxy, conf, cls in det_tidl:
                        xyxy = torch.tensor(xyxy).view(-1).tolist()
                        line = (conf, cls,  *xyxy) if opt.save_conf else (cls, *xyxy)  # label format
                        with open(txt_path + '.txt', 'a') as f:
                            f.write(('%g ' * len(line)).rstrip() % line + '\n')

            # Print time (inference + NMS)
            print(f'{s}Done. ({t2 - t1:.3f}s)')

            # Stream results
            if view_img:
                cv2.imshow(str(p), im0)
                cv2.waitKey(1)  # 1 millisecond

            # Save results (image with detections)
            if save_img:
                if dataset.mode == 'image':
                    cv2.imwrite(save_path, im0)
                else:  # 'video' or 'stream'
                    if vid_path != save_path:  # new video
                        vid_path = save_path
                        if isinstance(vid_writer, cv2.VideoWriter):
                            vid_writer.release()  # release previous video writer
                        if vid_cap:  # video
                            fps = vid_cap.get(cv2.CAP_PROP_FPS)
                            w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
                            h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
                        else:  # stream
                            fps, w, h = 30, im0.shape[1], im0.shape[0]
                            save_path += '.mp4'
                        vid_writer = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*'mp4v'), fps, (w, h))
                    vid_writer.write(im0)

    if save_txt or save_txt_tidl or save_img:
        s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt or save_txt_tidl else ''
        print(f"Results saved to {save_dir}{s}")

    print(f'Done. ({time.time() - t0:.3f}s)')


if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument('--weights', nargs='+', type=str, default='yolov5s.pt', help='model.pt path(s)')
    parser.add_argument('--source', type=str, default='data/images', help='source')  # file/folder, 0 for webcam
    parser.add_argument('--img-size', nargs= '+', type=int, default=640, help='inference size (pixels)')
    parser.add_argument('--conf-thres', type=float, default=0.01, help='object confidence threshold')
    parser.add_argument('--iou-thres', type=float, default=0.5, help='IOU threshold for NMS')
    parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
    parser.add_argument('--view-img', action='store_true', help='display results')
    parser.add_argument('--save-txt', action='store_true', help='save results to *.txt')
    parser.add_argument('--save-txt-tidl', action='store_true', help='save results to *.txt in tidl format')
    parser.add_argument('--save-bin', action='store_true', help='save base n/w outputs in raw bin format')
    parser.add_argument('--save-conf', action='store_true', help='save confidences in --save-txt labels')
    parser.add_argument('--save-crop', action='store_true', help='save cropped prediction boxes')
    parser.add_argument('--nosave', action='store_true', help='do not save images/videos')
    parser.add_argument('--classes', nargs='+', type=int, help='filter by class: --class 0, or --class 0 2 3')
    parser.add_argument('--agnostic-nms', action='store_true', help='class-agnostic NMS')
    parser.add_argument('--augment', action='store_true', help='augmented inference')
    parser.add_argument('--update', action='store_true', help='update all models')
    parser.add_argument('--project', default='runs/detect', help='save results to project/name')
    parser.add_argument('--name', default='exp', help='save results to project/name')
    parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
    parser.add_argument('--line-thickness', default=3, type=int, help='bounding box thickness (pixels)')
    parser.add_argument('--hide-labels', default=False, action='store_true', help='hide labels')
    parser.add_argument('--hide-conf', default=False, action='store_true', help='hide confidences')
    parser.add_argument('--kpt-label', type=int, default=5, help='number of keypoints')
    parser.add_argument("--widerface",action="store_true",help='save widerface_val txt')
    parser.add_argument('--save-widerface', type=str, default='./widerface_txt', help=' save widerface_txt folder') 
    opt = parser.parse_args()
    print(opt)
    check_requirements(exclude=('tensorboard', 'pycocotools', 'thop'))

    with torch.no_grad():
        if opt.update:  # update all models (to fix SourceChangeWarning)
            for opt.weights in ['yolov5s.pt', 'yolov5m.pt', 'yolov5l.pt', 'yolov5x.pt']:
                detect(opt=opt)
                strip_optimizer(opt.weights)
        else:
            detect(opt=opt)

执行命令

CUDA_VISIBLE_DEVICES=0 python detect-widerface.py  --weights your path  --source your input_path --widerface --save-widerface 2023-widerface-1600

3.最后一步就是将生成的txt文件移到指定的目录下，代码如下

import os
import os.path as osp
import re
import shutil
from pathlib import Path

if __name__ == "__main__":

	#用于评测的txt文件夹
    path = "./yolov7s_txt-pre-weight"
    #detect生成的txt文件
    source_path = "./2023-widerface-1600"
    dir_list = ["0--Parade","1--Handshaking","2--Demonstration","3--Riot","4--Dancing","5--Car_Accident","6--Funeral","7--Cheering","8--Election_Campain","9--Press_Conference","10--People_Marching","11--Meeting","12--Group","13--Interview","14--Traffic","15--Stock_Market","16--Award_Ceremony","17--Ceremony","18--Concerts","19--Couple","20--Family_Group","21--Festival","22--Picnic","23--Shoppers","24--Soldier_Firing","25--Soldier_Patrol","26--Soldier_Drilling","27--Spa","28--Sports_Fan","29--Students_Schoolkids","30--Surgeons","31--Waiter_Waitress","32--Worker_Laborer","33--Running","34--Baseball","35--Basketball","36--Football","37--Soccer","38--Tennis","39--Ice_Skating","40--Gymnastics","41--Swimming","42--Car_Racing","43--Row_Boat","44--Aerobics","45--Balloonist","46--Jockey","47--Matador_Bullfighter","48--Parachutist_Paratrooper","49--Greeting","50--Celebration_Or_Party","51--Dresses","52--Photographers","53--Raid","54--Rescue","55--Sports_Coach_Trainer","56--Voter","57--Angler","58--Hockey","59--people--driving--car","61--Street_Battle"]
    for dir_path in dir_list:
        obj_path = osp.join(path,dir_path)
        os.makedirs(obj_path,exist_ok=True)
    # num = 0
    # print("source_path===",source_path)
    for file_path in os.listdir(source_path):
        file_id_compile = re.compile(r"([\d]+)_")
        file_id = re.findall(file_id_compile,file_path)[0]
        file_paths = osp.join(source_path,file_path)
        
        dir_id_compile = re.compile(r"([\d]+)--")
        for path_dir in dir_list:
            dir_id = re.findall(dir_id_compile,path_dir)[0]
            # print("dir_id:%s"%(dir_id))
            if file_id == dir_id:
                # print("file_id===%s,dir_id===%s"%(file_id,dir_id))
                shutil.copyfile(file_paths,Path(Path(path)/path_dir)/file_path)
                break