Advertisement
Guest User

Untitled

a guest
Oct 16th, 2019
103
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 0.33 KB | None | 0 0
  1. Hyperparameter,value
  2. Number of Episodes, 2000
  3. Number of Timesteps, 1000
  4. Print Checkpoint step every, 4
  5. Training Batch Size, 64
  6. Discount Rate / Gamma, 0.99
  7. Learning Rate / alpha, 5e-4
  8. Number of Hidden Layers, 2
  9. Fully Connected Layer 1 Units, 64
  10. Fully Connected Layer 2 Units, 64
  11. TAU, 1e-3
  12. Epsilon, 0.1
  13. Epsilon-Min, 0.01
  14. Epsilon-Decay, 0.995
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement