I notice that after the train ceased, I can't resume training based on my half-tained model.
Is it because the parameter RESUME is set false in the DOTA_quadrangle.yaml file?
If I set it true , will it automatically resume training based on the params and states file I got in the last training?
I notice that after the train ceased, I can't resume training based on my half-tained model. Is it because the parameter RESUME is set false in the DOTA_quadrangle.yaml file? If I set it true , will it automatically resume training based on the params and states file I got in the last training?