Closed purudpd closed 6 years ago
Hi @purudpd , I got very similar test output to yours and have concluded that the expected results for test cases probably need to be adjusted because the updated underlying libraries (probably dlib?) are producing slightly different results. I didn't feel comfortable doing this in my pull request so have left @bamos to look at it, being his project - that seems to be his prerogative.
See #172 and associated pull request for more detail.
All of my test cases fail after successfully installing open-face and all of its dependencies as given http://cmusatyalab.github.io/openface/setup/
I had installed the latest version of both Open-CV and DLIB. OpenCV -> ver_3.2.0-dev DLIB -> ver_19.4.0
Please see this line in output below. UserWarning: Trying to unpickle estimator SVC from version pre-0.18 when using version 0.18.1. This might lead to breaking code or invalid results. Use at your own risk. (Could the issue be related to the sklearn library version?)
Here is my output from the tests.