Closed AndyHon closed 4 years ago
I have uploaded the files now, and modified some code in train.py and test.py . The object_detection is borrowed from tensorflow/models. @AndyHon
@fjchange Thanks for your timely reply. There should be “mscoco_label_map.pbtxt” label file in folder object_detection/data/,Could you please upload it, thank you.
In inference.py: the line of 159
np_paths_boxes_path = '/home/'+args.machine+'/'+args.dataset+'_'+'img_path_box.npy'
args.machine shuold be delete if dataset is avenue, But I don't know what 'img_path_box.npy' is. Could you tell me how to get this file? I'm sorry to bother you so many times.
The file have been uploaded. The file_path is set machine related ( I train it in different servers ) I will update the code to make it independence.
@fjchange from object_detection.core import standard_fields as fields Thanks for the update,but object_detection file should lack of "core" file
object_detection.utils import static_shape object_detection.utils file should lack static_shape.py file Thank you for constantly updating the content. Could you please run the experiment again to see what files or codes are missing in the whole document? Then upload your project in its entirety. thank you very much.
Hello, do you have time to upload the missing code?
Code should be completed now.
from ._conv import register_converters as _register_converters Traceback (most recent call last): File "/media/fxc/KESU/object_centric_VAD/scripts/inference.py", line 9, in <module> from object_detection.utils import visualization_utils as vis_util File "../object_detection/utils/visualization_utils.py", line 42, in <module> from object_detection.utils import shape_utils File "../object_detection/utils/shape_utils.py", line 25, in <module> from object_detection.utils import static_shape ImportError: cannot import name 'static_shape'
@fjchange Sorry to bother you again. You should be lack a file static_shape.py in /object_detection/ utils
Can you upload complete files and programs in object_detection folder?I would appreciate it very much.