I know most of you who have worked with tensorflow object detection have come across this problem. I have my tensorflow project running in a different folder but now I did like to use object detection functions stored in models folder.
I have changed my working directory in my python scripts to models / object detection folder this way:
import sys
import os
changed_directory = False
if not changed_directory:
os.chdir(os.path.join('../models/research/','object_detection/'))
changed_directory = True
I then append the path in my scripts this way :
import sys
sys.path.append("...models\\research\\object_detection\\")
This is how my models folder looks like
TensorFlow/
NAR_object_detection/
└─ models/
├─ community/
├─ official/
├─ orbit/
├─ research/
└── ...
Object detection folder in inside the research models. Also important to say that my python scripts for object detection are stored in a folder called NAR_object_detection inside the Tensorflow, so I mean NAR_object_detection folder is the same folder with models folder.
When I try to import this way I get an error that no module called object detection
from object_detection.utils import label_map_util
Why is it that the import statement is not finding the object detection folder?
CodePudding user response:
try this
pip install tensorflow-object-detection-api