MMselfSup自监督预训练模型的评估：“检测”下游任务

1.提取自监督预训练模型的 backbone 权值

python tools/model_converters/extract_backbone_weights.py {CHECKPOINT} {MODEL_FILE} 参数:
CHECKPOINT：自监督预训练过程中保存下来（名为epoch_*.pth）的模型文件路径
MODEL_FILE：输出 backbone 权重文件的保存路径。

 生成的 backbone 权值文件

2. 配置环境

创建虚拟环境

conda create --name openmmlab python=3.8 -y

激活虚拟环境：

conda activate openmmlab

安装pytorch、torchvision

根据自己的配置安装相应版本

pip install torch==1.7.1+cu101 torchvision==0.8.2+cu101 -f https://download.pytorch.org/whl/torch_stable.html

或手动下载，地址：https://download.pytorch.org/whl/torch_stable.html

下载I MMEngine 和 MMCV

pip install -U openmim
mim install mmengine
mim install 'mmcv>=2.0.0rc1'
注：推荐使用命令：pip install mmcv==2.0.0rc1 -f https://download.openmmlab.com/mmcv/dist/cu101/torch1.7/index.html进行下载

安装mmdetection3.0版本

git clone https://github.com/open-mmlab/mmdetection.git

cd mmdetection
pip install -v -e .

3.写目标检测任务的配置文件

在路径：/configs/faster_rcnn/下

新建一个名为faster_rcnn_simclr-pretrained_r50_fpn_9k_coco.py  的配置文件。

写入内容

_base_ = 'faster_rcnn_r50_fpn_90k_coco.py'
model = dict(
    backbone=dict(
        frozen_stages=-1,
        init_cfg=dict(
             type='Pretrained',
            checkpoint='checkpoints/200.pth')
 ))
# optimizer
optimizer = dict(
    lr=0.02 * (1 / 8))
# Runner type
runner = dict(_delete_=True, type='IterBasedRunner', max_iters=9000)
checkpoint_config = dict(interval=3000)
evaluation = dict(interval=3000)

4.修改coco_detection.py.py文件

主要修改coco数据集路径

 

 5.修改coco.py

主要修改CLASSES

 6.修改基础backbones网络

主要修改num_classes

 7.训练

8.测试

使用tool/test.py

 

 结果

修改image_demo.py可进行批量检测，代码如下：

import asyncio
from argparse import ArgumentParser
import mmcv
from mmdet.apis import (async_inference_detector, inference_detector,
                        init_detector)
from mmdet.registry import VISUALIZERS
from mmdet.utils import register_all_modules
import os
import cv2
def parse_args():
    parser = ArgumentParser()
    parser.add_argument('img', help='Image file')
    parser.add_argument('config', help='Config file')
    parser.add_argument('checkpoint', help='Checkpoint file')
    parser.add_argument('--out-file', default=None, help='Path to output file')
    parser.add_argument(
        '--device', default='cuda:0', help='Device used for inference')
     parser.add_argument(
        '--palette',
        default='coco',
        choices=['coco', 'voc', 'citys', 'random'],
        help='Color palette used for visualization')
    parser.add_argument(
        '--score-thr', type=float, default=0.3, help='bbox score threshold')
    parser.add_argument(
        '--async-test',
        action='store_true',
        help='whether to set async options for async inference.')
    args = parser.parse_args()
    return args
def main(args):
    register_all_modules()
    file_name = os.listdir(args.img)
    model = init_detector(
        args.config, args.checkpoint,palette=args.palette, device=args.device)
    visualizer = VISUALIZERS.build(model.cfg.visualizer)
    visualizer.dataset_meta = model.dataset_meta
    for images in file_name:
        name = images
        images = os.path.join(args.img, images)
        result = inference_detector(model, images)
        img = mmcv.imread(images)
        img = mmcv.imconvert(img, 'bgr', 'rgb')
        visualizer.add_datasample(
            name,
            img,
            data_sample=result,
            draw_gt=False,
            show=args.out_file is None,
            wait_time=0,
            out_file=args.out_file + name,
            pred_score_thr=args.score_thr)
async def async_main(args):
    model = init_detector(args.config, args.checkpoint, device=args.device)
    visualizer = VISUALIZERS.build(model.cfg.visualizer)
    visualizer.dataset_meta = model.dataset_meta
    tasks = asyncio.create_task(async_inference_detector(model, args.img))
    result = await asyncio.gather(tasks)
    img = mmcv.imread(args.img)
    img = mmcv.imconvert(img, 'bgr', 'rgb')
    visualizer.add_datasample(
        'result',
        img,
        pred_sample=result[0],
        show=args.out_file is None,
        wait_time=0,
        out_file=args.out_file,
        pred_score_thr=args.score_thr,
 )
if __name__ == '__main__':
    args = parse_args()
    assert not args.async_test, 'async inference is not supported yet.'
    if args.async_test:
        asyncio.run(async_main(args))
    else:
        main(args)
          

 运行：

 

 批量测试结果：