Betapose: Estimating 6D Pose From Localizing Designated Surface Keypoints

上交这份code质量堪忧啊,缺斤少两的还一堆bug,文档写的也是愁死人。等我配通了再来填坑。

1.from utils import output_pointcloud ImportError: cannot import name 'output_pointcloud'

需要修改utils/model.py

from utils import output_pointcloud

改为:
 

from utils.utils import output_pointcloud

2.Traceback (most recent call last): File "src/train.py", line 14, in from evaluation import prediction File "...\betapose\3_6Dpose_estimator\train_KPD\src\evaluation.py", line 9, in from predict.annot.coco_minival import Mscoco_minival ModuleNotFoundError: No module named 'predict.annot'

删去这一行:

 from predict.annot.coco_minival import Mscoco_minival 

3. 

  os.mkdir("../exp/{}/{}".format(opt.dataset, opt.expID))
FileNotFoundError: [Errno 2] No such file or directory: '../exp/coco/seq5_Nov_1_1'
  File "src/train.py", line 129, in main
    os.mkdir("../exp/{}".format(opt.dataset))
FileNotFoundError: [Errno 2] No such file or directory: '../exp/coco'

在betapose/3_6Dpose_esitimator目录下面创建exp/coco文件夹

 

 

4.OSError: Unable to open file (unable to open file: name = './data/05/annot_train.h5', errno = 2, error message = 'No such file or directory', flags = 0, o_flags = 0)

在betapose/3_6Dpose_esitimator/trian_KPD/下创建data/05/并将步骤二中产生的annot_train.h5,annot_eval.h5,trian等一堆文件复制到这里

 

5.昨天又解决一个问题,老师亲自跑了下才弄懂为啥。

Prepare data as AlexeyAB/darknet's instructions. Refer to folder ./scripts for more help

这句话本来一直没看懂作者的意思,卡了快两个星期了,issues也没人回。原来就是跑一下./scripts文件夹下的gt_single_object.py,然后也不能直接运行Run train_single.sh。打开train_single.sh,把命令改成

./darknet detector train data_linemod_gt/02/bvise.data cfg/yolo-linemod-single.cfg backup_linemod_no_neg/02/darknet53.conv.74

backup_linemod_no_neg/02/这个文件夹自己建一下,darknet53.conv.74作者给了地址下载。data_linemod_gt/02/bvise.data这个是gt_single_object.py跑通之后生成的。自己建文件夹bvise.data扔进去。还没完,建一个backup_linemod_gt/02文件夹保存结果。醉了,这么多事,作者就这么几句话一带而过.....

你可能感兴趣的:(AR深度学习项目)