simoninithomas / Deep_reinforcement_learning_Course

Implementations from the free course Deep Reinforcement Learning with Tensorflow and PyTorch
http://www.simoninithomas.com/deep-rl-course
3.74k stars 1.23k forks source link

Process got killed after Episode 17 #62

Open ravishk1 opened 5 years ago

ravishk1 commented 5 years ago

I was training the same code that you have provided, But it is getting killed after 17 episodes. What can I infer from this? This happened 3 times!

Model Saved Episode: 1 Total reward: 105.0 Explore P: 0.9547 Training Loss 0.0046 Episode: 2 Total reward: 105.0 Explore P: 0.9408 Training Loss 0.0009 Episode: 3 Total reward: 80.0 Explore P: 0.9239 Training Loss 0.0023 Episode: 4 Total reward: 215.0 Explore P: 0.8979 Training Loss 6.2312 Episode: 5 Total reward: 460.0 Explore P: 0.8730 Training Loss 0.0272 Model Saved Episode: 6 Total reward: 255.0 Explore P: 0.8475 Training Loss 0.0019 Episode: 7 Total reward: 315.0 Explore P: 0.8226 Training Loss 1.5508 Episode: 8 Total reward: 210.0 Explore P: 0.8038 Training Loss 0.0031 Episode: 9 Total reward: 365.0 Explore P: 0.7788 Training Loss 0.0116 Episode: 10 Total reward: 260.0 Explore P: 0.7584 Training Loss 0.0428 Model Saved Episode: 11 Total reward: 510.0 Explore P: 0.7363 Training Loss 1.5399 Episode: 12 Total reward: 210.0 Explore P: 0.7190 Training Loss 0.0216 Episode: 13 Total reward: 215.0 Explore P: 0.7002 Training Loss 1.9142 Episode: 14 Total reward: 210.0 Explore P: 0.6834 Training Loss 0.3903 Episode: 15 Total reward: 415.0 Explore P: 0.6601 Training Loss 1.5385 Model Saved Episode: 16 Total reward: 285.0 Explore P: 0.6416 Training Loss 0.0108 Episode: 17 Total reward: 180.0 Explore P: 0.6230 Training Loss 0.0026 Killed

pengzhi1998 commented 5 years ago

Yes, I think it is because your memory is not enough, this kind of training does need a lot of resource