Skip to content

Feature ๐Ÿค–

  • ๐Ÿ‘จโ€๐Ÿ‘ฉโ€๐Ÿ‘งโ€๐Ÿ‘ฆ Unified code structure
  • ๐Ÿ“„ Single-file implementation
  • ๐Ÿท Low code reuse
  • ๐Ÿ“ Minimizing code differences
  • ๐Ÿ“ˆ Tensorboard & Wandb integration
  • ๐Ÿ›ค PEP8(code style) & PEP526(type hint) compliant

๐Ÿ“ Minimizing code differences

In order to facilitate the comparison of the differences between algorithms and to unify the code style, the code will try to minimize the differences between the wired code files as shown in the diagram below. 

graph LR
A[dqn_torch.py] -->B[ddpg_torch.py];
B -->C[td3_torch.py];
C -->D[sac_torch.py];
B -->E[ppo_torch.py];
A -->F[ddqn_torch.py];
A -->G[pdqn_torch.py];
A -->H[dqn_atari_torch.py];
A -->I[dqn_tf.py];
F -->J[ddqn_tf.py];
G -->K[pdqn_tf.py];
H -->L[dqn_atari_tf.py];

Last update: 2023-02-21