First model, first training

Can anyone please help me with understanding the training process. This is the first ever model I train. I'm renting GPU online and it's been going for 3 days, completed 500/1000 episodes which is about 7.5 million steps. However by looking at the rewards lines plots I cannot tell if it's learning anything. It's reinforcement learning using A3C agent. Is it worth ending the training and then test it, or there is no point?

deeplearning episodes gpu plots process reinforcement reinforcement learning training understanding