Replies: 2 comments 2 replies
-
Hi, it should be able to complete the |
Beta Was this translation helpful? Give feedback.
0 replies
-
Hi again. Here is a pickled output of the import pickle as pkl
import cv2
with open("screenshot.pkl", "rb") as f:
snap = pkl.load(f)
print(snap.shape)
cv2.imshow("test", snap)
cv2.waitKey(0)
cv2.destroyAllWindows() |
Beta Was this translation helpful? Give feedback.
2 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hi, I am working on an extension to enable the RolloutWorkers run on Linux. I have an interface that let's me run and steer Trackmania in Lutris, but when I run the provided agent on the test-track, with the given weights, the car is not able to finish the track.
It could be that the CNN is confused, as I am not sure what the function WindowInterface.screenshot() is returning. I had to make assumptions (returns with default settings a 128x256x4 np array, RGB and opacity), since I can't confirm it on a Windows machine. I hope you can help me out with following two questions:
q1: With the given weights and the test setup, the car should be able to succeed?
q2: what is the format of img that is being returned by screenshot? (np.array of dim ..., whereas arr[0] is ...)
Thanks a lot,
p
Beta Was this translation helpful? Give feedback.
All reactions