mmdetection 安装与使用（win10）

我会带着你远行 2023-07-01 04:23 19阅读 0赞

一、安装

（1）创建虚拟环境：

conda create -n mmdec python=3.6

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70][]

完成

（2）激活虚拟环境，安装torch（现在mmdetection需要的torch最低版本为1.1.0）

activate mmdec

发现官网的这个pip也是挺快的（[https://pytorch.org/][https_pytorch.org]）

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 1][]

pip install http://download.pytorch.org/whl/cu100/torch-1.1.0-cp36-cp36m-win\_amd64.whl （本人使用）

pip (--default-timeout=600等待时间) install \*\*\*\*\*\* 若果超时可以用下面的多等待一会儿

其他版本路径：

https://download.pytorch.org/whl/cpu/torch-1.1.0-cp36-cp36m-win_amd64.whl（WIN,CPU）
    
    https://download.pytorch.org/whl/cu100/torch-1.1.0-cp36-cp36m-win_amd64.whl （WIN,CUDA 10.0）
    https://download.pytorch.org/whl/cu90/torch-1.1.0-cp36-cp36m-win_amd64.whl （WIN,CUDA 9.0）
    
    pip install https://download.pytorch.org/whl/cu100/torch-1.0.1-cp35-cp35m-linux_x86_64.whl（linux，CUDA 10.0）

pip install PyYAML -i https://pypi.tuna.tsinghua.edu.cn/simple

pip install cython

pip install scikit-image

pip install torchvision

Pycocotools 这个东西windows安装需要[https://github.com/philferriere/cocoapi][https_github.com_philferriere_cocoapi]下载后进入pythonAPI文件夹内

使用python setup.py install

继续安装上面缺少的库：

pip install 库名称 -i [https://pypi.tuna.tsinghua.edu.cn/simple][https_pypi.tuna.tsinghua.edu.cn_simple]

V2EX：[http://pypi.v2ex.com/simple][http_pypi.v2ex.com_simple]  
豆瓣：[http://pypi.douban.com/simple][http_pypi.douban.com_simple]  
中国科学技术大学：[http://pypi.mirrors.ustc.edu.cn/simple][http_pypi.mirrors.ustc.edu.cn_simple]  
清华：[https://pypi.tuna.tsinghua.edu.cn/simple][https_pypi.tuna.tsinghua.edu.cn_simple]

(3)安装mmcv

**直接pip安装就可以：pip install mmcv **

**以前 **我是按照以下安装的

下载：[https://github.com/open-mmlab/mmcv][https_github.com_open-mmlab_mmcv]

pip install .

补充，电脑重装遇到的问题：anaconda3\\envs\\torch\\lib\\site-packages\\pip\\compat\\\_\_init\_\_.py", line 75, in console\_to\_str  
    return s.decode('utf\_8')  
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xa1 in position 60: invalid start byte

发现是pip版本过低，更新pip 解决问题| python -m pip install --upgrade pip

（4）安装

**mmdetection 1.x版本**：如果你安装的是VS2015，自定义安装，updata3需要选上，否则这一步失败。

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 2][]

下载：[https://github.com/open-mmlab/mmdetection][https_github.com_open-mmlab_mmdetection]

使用如下命令安装

python setup.py develop -i https://pypi.tuna.tsinghua.edu.cn/simple

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 3][]

**mmdetection 2.x版本**，使用原本的VS2015报错，然后我安装了vs2019结果cuda又不支持了，我太难了

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 4][]

纠结了一下是升级cuda还是卸载vs2019，决定重新安装一个VS2017 试试。

待续。。。2.1版本的报错，先安装了个2.0 的

的![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 5][]

ubuntu下和1.x一样的安装方式正常安装

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 6][]

（5）测试

from mmdet.apis import init_detector, inference_detector, show_result
    config_file = './configs/faster_rcnn_r50_fpn_1x.py'
    checkpoint_file = './checkpoints/faster_rcnn_r50_fpn_1x_20181010-3d1b3351.pth'
    model = init_detector(config_file, checkpoint_file)
    img = 'demo.jpg'
    result = inference_detector(model, img)
    show_result(img, result, model.CLASSES)

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 7][]

遇到问题：

**0、ImportError: cannot import name 'get\_dist\_info'**

mmcv版本和mmdetection版本不匹配

1、在测试的时候提示：**AttributeError: module 'torch.distributed' has no attribute 'is\_initialized'**定位到错误处修改

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 8][]

\-------------------------------------------------------分割线-------------------------------------------------

使用自己的数据集进行训练

1数据集准备：

因为一直用的都是VOC数据集，所以直接训练Voc的数据集

2.在mmdet/datasets文件夹内，修改自己的数据

1、在\_\_init\_\_ 中新增自己的

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 9][]

2、复制voc重命名成自己的MyData.py修改地方已标红

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 10][]

3.修改configs中的配置文件

# model settings
    model = dict(
        type='FasterRCNN',
        pretrained='torchvision://resnet50',
        backbone=dict(
            type='ResNet',
            depth=50,
            num_stages=4,
            out_indices=(0, 1, 2, 3),
            frozen_stages=1,
            style='pytorch'),
        neck=dict(
            type='FPN',
            in_channels=[256, 512, 1024, 2048],
            out_channels=256,
            num_outs=5),
        rpn_head=dict(
            type='RPNHead',
            in_channels=256,
            feat_channels=256,
            anchor_scales=[8],
            anchor_ratios=[0.5, 1.0, 2.0],
            anchor_strides=[4, 8, 16, 32, 64],
            target_means=[.0, .0, .0, .0],
            target_stds=[1.0, 1.0, 1.0, 1.0],
            loss_cls=dict(
                type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
            loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
        bbox_roi_extractor=dict(
            type='SingleRoIExtractor',
            roi_layer=dict(type='RoIAlign', out_size=7, sample_num=2),
            out_channels=256,
            featmap_strides=[4, 8, 16, 32]),
        bbox_head=dict(
            type='SharedFCBBoxHead',
            num_fcs=2,
            in_channels=256,
            fc_out_channels=1024,
            roi_feat_size=7,
            num_classes=3,  # 类别+1
            target_means=[0., 0., 0., 0.],
            target_stds=[0.1, 0.1, 0.2, 0.2],
            reg_class_agnostic=False,
            loss_cls=dict(
                type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0),
            loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0)))
    # model training and testing settings
    train_cfg = dict(
        rpn=dict(
            assigner=dict(
                type='MaxIoUAssigner',
                pos_iou_thr=0.7,
                neg_iou_thr=0.3,
                min_pos_iou=0.3,
                ignore_iof_thr=-1),
            sampler=dict(
                type='RandomSampler',
                num=256,
                pos_fraction=0.5,
                neg_pos_ub=-1,
                add_gt_as_proposals=False),
            allowed_border=0,
            pos_weight=-1,
            debug=False),
        rpn_proposal=dict(
            nms_across_levels=False,
            nms_pre=2000,
            nms_post=2000,
            max_num=2000,
            nms_thr=0.7,
            min_bbox_size=0),
        rcnn=dict(
            assigner=dict(
                type='MaxIoUAssigner',
                pos_iou_thr=0.5,
                neg_iou_thr=0.5,
                min_pos_iou=0.5,
                ignore_iof_thr=-1),
            sampler=dict(
                type='RandomSampler',
                num=512,
                pos_fraction=0.25,
                neg_pos_ub=-1,
                add_gt_as_proposals=True),
            pos_weight=-1,
            debug=False))
    test_cfg = dict(
        rpn=dict(
            nms_across_levels=False,
            nms_pre=1000,
            nms_post=1000,
            max_num=1000,
            nms_thr=0.7,
            min_bbox_size=0),
        rcnn=dict(
            score_thr=0.05, nms=dict(type='nms', iou_thr=0.5), max_per_img=100)
        # soft-nms is also supported for rcnn testing
        # e.g., nms=dict(type='soft_nms', iou_thr=0.5, min_score=0.05)
    )
    # dataset settings
    dataset_type = 'MyData'
    data_root = 'E:/mm/VOCdevkit/'
    img_norm_cfg = dict(mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
    data = dict(
        imgs_per_gpu=2,
        workers_per_gpu=2,
        train=dict(
            type='RepeatDataset',  # to avoid reloading datasets frequently
            times=3,
            dataset=dict(
                type=dataset_type,
                ann_file=[data_root + 'VOC2007/ImageSets/Main/train.txt'],
                img_prefix=[data_root + 'VOC2007/'],
                img_scale=(400, 400),
                img_norm_cfg=img_norm_cfg,
                size_divisor=32,
                flip_ratio=0.5,
                with_mask=False,
                with_crowd=False,
                with_label=True)),
        val=dict(
            type=dataset_type,
            ann_file=data_root + 'VOC2007/ImageSets/Main/val.txt',
            img_prefix=data_root + 'VOC2007/',
            img_scale=(400, 400),
            img_norm_cfg=img_norm_cfg,
            size_divisor=32,
            flip_ratio=0,
            with_mask=False,
            with_crowd=False,
            with_label=True),
        test=dict(
            type=dataset_type,
            ann_file=data_root + 'VOC2007/ImageSets/Main/test.txt',
            img_prefix=data_root + 'VOC2007/',
            img_scale=(400, 400),
            img_norm_cfg=img_norm_cfg,
            size_divisor=32,
            flip_ratio=0,
            with_mask=False,
            with_crowd=False,
            with_label=False,
            test_mode=True))
    
    # optimizer
    optimizer = dict(type='SGD', lr=0.02, momentum=0.9, weight_decay=0.0001)
    optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
    # learning policy
    lr_config = dict(
        policy='step',
        warmup='linear',
        warmup_iters=500,
        warmup_ratio=1.0 / 3,
        step=[8, 11])
    checkpoint_config = dict(interval=1)
    # yapf:disable
    log_config = dict(
        interval=100,
        hooks=[
            dict(type='TextLoggerHook'),
            # dict(type='TensorboardLoggerHook')
        ])
    # yapf:enable
    # runtime settings
    total_epochs = 12
    dist_params = dict(backend='nccl')
    log_level = 'INFO'
    work_dir = 'model'
    load_from = None
    resume_from = None
    workflow = [('train', 1)]

4.使用tools中的train训练

config 变为--config =“配置文件”

5.测试

下载权重

常见问题（本人遇见的）：

1、在训练的时候提示的：KeyError: 'None is not in the dataset registry'

修改mmdet/datasets内修改不完整

2、在训练完一个epoch保存权重后提示的：OSError: symbolic link privilege not held

以管理员身份运行pycharm

![20200119090124622.png][]

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70]: https://img-blog.csdnimg.cn/20200119085959454.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg==,size_16,color_FFFFFF,t_70
[https_pytorch.org]: https://pytorch.org/
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 1]: https://img-blog.csdnimg.cn/20200414204030106.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg==,size_16,color_FFFFFF,t_70
[https_github.com_philferriere_cocoapi]: https://github.com/philferriere/cocoapi
[https_pypi.tuna.tsinghua.edu.cn_simple]: https://link.jianshu.com/?t=https:/pypi.tuna.tsinghua.edu.cn/simple
[http_pypi.v2ex.com_simple]: https://link.jianshu.com/?t=http:/pypi.v2ex.com/simple
[http_pypi.douban.com_simple]: https://link.jianshu.com/?t=http:/pypi.douban.com/simple
[http_pypi.mirrors.ustc.edu.cn_simple]: https://link.jianshu.com/?t=http:/pypi.mirrors.ustc.edu.cn/simple/
[https_github.com_open-mmlab_mmcv]: https://github.com/open-mmlab/mmcv
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 2]: https://img-blog.csdnimg.cn/20200608125753606.jpg?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg==,size_16,color_FFFFFF,t_70
[https_github.com_open-mmlab_mmdetection]: https://github.com/open-mmlab/mmdetection
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 3]: https://img-blog.csdnimg.cn/20200119101326649.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 4]: https://img-blog.csdnimg.cn/20200615193025295.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 5]: https://img-blog.csdnimg.cn/20200616132948795.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 6]: https://img-blog.csdnimg.cn/20200616100712932.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 7]: https://img-blog.csdnimg.cn/20200119102016428.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 8]: https://img-blog.csdnimg.cn/20200119102147476.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 9]: https://img-blog.csdnimg.cn/20200119083027156.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg_size_16_color_FFFFFF_t_70 10]: https://img-blog.csdnimg.cn/20200119083429706.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1RZdGFuZ3lhbg==,size_16,color_FFFFFF,t_70
[20200119090124622.png]: https://img-blog.csdnimg.cn/20200119090124622.png