I'm sorry it's taken me a full week to respond - lot's of stuff has been going on recently and it's been year since I last worked on this so I'm a bit rusty.
In your original comment, I saw you changed INPUT_SHAPE, but did you also change the process_frame function in train_dqn.py to not grayscale the image? IIRC, you can make it not grayscale the image by removing frame = cv2.cvtColor(frame, cv2.COLOR_RGB2GRAY)
I'm sorry it's taken me a full week to respond - lot's of stuff has been going on recently and it's been year since I last worked on this so I'm a bit rusty.
In your original comment, I saw you changed
INPUT_SHAPE, but did you also change theprocess_framefunction intrain_dqn.pyto not grayscale the image? IIRC, you can make it not grayscale the image by removingframe = cv2.cvtColor(frame, cv2.COLOR_RGB2GRAY)