I'm working on a single animal project in DeepLabCut and everytime I try to train the network i get an error "ValueError: Image size [2160 3840] may be too large. Consider increasing max_input_size and/or decreasing global_scale in the train/pose_cfg.yaml."
Changed the parameter value in the pose file, no issues saving the change in scale but it still keeps resetting! I can't seem to find a code block to change the global_scale either. No similar parameter found in the config file.
As suggested,I tried changing the global_scale in the pose_cfg.yaml (from 0.8 to 0.5), saved it and ran the program again, the same error popped up and when I checked the file the value has reset to 0.5 I tried this 2-3 times to make sure I was saving it correctly, and as expected the value did save and update in the pose file but still kept resetting everytime I ran the train function.
I'm trying this in python-3 on colab.
