search

sergiomsilva / alpr-unconstrained

License Plate Detection and Recognition in Unconstrained Scenarios
Other
1.71k stars 604 forks source link

what environment did you guys use? #131

Open chen3082 opened 4 years ago

chen3082 commented 4 years ago

I tried to copy Ubuntu 16.04 machine, with Keras 2.2.4, TensorFlow 1.5.0, OpenCV 2.4.9, NumPy 1.14 and Python 2.7 or Python3.5. However, I encountered a lot of problems.

First, installing OpenCV2.4.9 is a pain. I'm not even sure did I do this correctly(I downloaded the zip file). How did you guys do it? I tried to import cv2 in python2.7. It shows "ImportError: numpy.core.multiarray failed to import”

Second, although I set the default python as 2.7 by alias python=python2 at the top of the ~/.bashrc . Every time I tried to run "run.sh", it kept showing missing parentheses for "print". And every time I tried "python --version" in the terminal, it shows python2.7

Third, I tried to install "yolo34py" as python wrapper by "pip install yolo34py". It also shows error.

I'm wondering how you guys set up the environment. Any help would be appreciated!!!! Thanks!

ulanbator commented 4 years ago

In ubuntu18.04,I can run this project successfully.this is what I have done conda create -n alpr-unconstrained python=2.7 conda activate alpr-unconstrained pip install keras==2.2.4 pip install tensorflow==1.5.0 conda install -c menpo opencv

chen3082 commented 4 years ago

@ulanbator Thank you sir :)