#stable baselines

[light606f]

NatureのDQNのCNN層

NatureのDQN:ttps://www.nature.com/articles/nature14236

これのCNN層はこんな構成

Stable baselines の CnnPolicy と多分一緒．

https://elix-tech.github.io/ja/2016/06/29/dqn-ja.html

[light606f]

Name Refactored [1] Recurrent Box Discrete Multi Processing A2C ✔️ ✔️ ✔️ ✔️ ✔️ ACER ✔️ ✔️ ❌ [4] ✔️ ✔️ ACKTR ✔️ ✔️ ✔️ ✔️ ✔️ DDPG ✔️ ❌ ✔️ ❌ ✔️ [3] DQN ✔️ ❌ ❌ ✔️ ❌ HER ✔️ ❌ ✔️ ✔️ ❌ GAIL [2] ✔️ ✔️ ✔️ ✔️ ✔️ [3] PPO1 ✔️ ❌ ✔️ ✔️ ✔️ [3] PPO2 ✔️ ✔️ ✔️ ✔️ ✔️ SAC ✔️ ❌ ✔️ ❌ ❌ TD3 ✔️ ❌ ✔️ ❌ ❌ TRPO ✔️ ❌ ✔️ ✔ ✔️ [3]

Name Name unomitted Year URL Detail 派生元 A3C Asynchronous Advantage Actor-Critic 2016 A2C Advantage Actor-Critic 2016 A3C ACER Actor-Critic with Experience Replay 2016 A3Cを方策オフ型に書き換えて, 経験再生を利用するアルゴリズムです. A3C ACKTR Actor Critic using Kronecker-Factored Trust Region 2017 PPO?; Actor-CritC,TRPO,Kronecker factorization DDPG Deep DPG 2015 DPG DQN 2013 https://arxiv.org/abs/1312.5602 HER Hindsight Experience Replay 2017 DDPG? GAIL Generative Adversarial Imitation Learning 2016 TRPO? PPO Proximal Policy Optimization 2017 https://openai.com/blog/openai-baselines-ppo/ TRPO SAC Soft Actor-Critic 2018 https://arxiv.org/abs/1801.01290 TD3 TD3 Twin Delayed DDPG 2018 https://arxiv.org/abs/1802.09477 DDPG TRPO 2015 https://arxiv.org/abs/1502.05477 DQN?; VPG?

https://qiita.com/namakemono/items/cfd4db78e2fec738a65c

https://qiita.com/shionhonda/items/ec05aade07b5bea78081