Open khany27 opened 3 years ago
I'm having the same problem here. Running the docker installation, and then attempting to run model_main_tf2.py from inside the container gives me this error:
Traceback (most recent call last):
File "object_detection/model_main_tf2.py", line 32, in <module>
from object_detection import model_lib_v2
File "/home/tensorflow/.local/lib/python3.6/site-packages/object_detection/model_lib_v2.py", line 29, in <module>
from object_detection import eval_util
File "/home/tensorflow/.local/lib/python3.6/site-packages/object_detection/eval_util.py", line 36, in <module>
from object_detection.metrics import lvis_evaluation
File "/home/tensorflow/.local/lib/python3.6/site-packages/object_detection/metrics/lvis_evaluation.py", line 23, in <module>
from lvis import results as lvis_results
File "/home/tensorflow/.local/lib/python3.6/site-packages/lvis/__init__.py", line 5, in <module>
from lvis.vis import LVISVis
File "/home/tensorflow/.local/lib/python3.6/site-packages/lvis/vis.py", line 1, in <module>
import cv2
File "/home/tensorflow/.local/lib/python3.6/site-packages/cv2/__init__.py", line 5, in <module>
from .cv2 import *
ImportError: libGL.so.1: cannot open shared object file: No such file or directory
I had the same issue.
My workaround was installing opencv-python-headless==4.2.0.34 with:
pip install opencv-python-headless==4.2.0.34
instead and it found the files.
It didn't work with versions 4.3.x.x+ so there may have been a change? I couldn't find any open issues regarding this at opencv-pyhton's GitHub repository at first glance.
@double-em Thanks!
I don't understand the package chain going on here, but it also seems to work for me to add opencv-python-headless==4.2.0.34
to the REQUIRED_PACKAGES
list in setup.py
.
The setup.py script installs the false version of packages and rechanging it to Tensorflow 2.2 makes it worse. I can run the inference of pre-trained models but cannot train on custom data. the model_main_tf2.py gives an error always. Please check the packages which are compatible with the Tensorflow version.