267 Commits
Author SHA1 Message Date
Mike Clark ed943a6356 Use logits in CategoricalActorCriticNet
A really minor change. Should we input logits instead of probs into `Categorical`? It will do less converting from logits and back, leading to increased stability during training (theoretically).
2018-06-15 08:24:14 +08:00
Shangtong Zhang a7e070077d Fix a bug of Huber loss 2018-06-12 21:52:15 -06:00
Shangtong Zhang 19572d0640 Fix a bug in deterministic evaluation 2018-05-30 14:53:08 -06:00
Shangtong Zhang 14c04eff87 Update plotter 2018-05-30 11:37:13 -06:00
Shangtong Zhang 66f5c37407 Update DDPG params and evaluation protocol 2018-05-30 11:24:23 -06:00
Shangtong Zhang 5a1d489938 Deterministic episodes 2018-05-29 11:02:37 -06:00
Shangtong Zhang 89b1450df7 Force random seed 2018-05-28 15:27:42 -06:00
Shangtong Zhang 5e91086c17 Fix a bug for single process 2018-05-25 19:52:58 -06:00
Shangtong Zhang d7c537707a Fix a minor bug of option critic 2018-05-20 21:56:14 -06:00
Shangtong Zhang 7066fb8e89 Import pybullet_envs 2018-05-19 22:10:32 -06:00
Shangtong Zhang 98d7a8a482 Update image 2018-05-19 11:41:07 -06:00
Shangtong Zhang 28a9e41a08 Remove tf logger 2018-05-19 11:31:56 -06:00
Shangtong Zhang d846bc50d7 Update version 2018-05-19 09:47:15 -06:00
Shangtong Zhang 5cfb5be18b Add figure for DDPG with pixels 2018-05-19 09:36:59 -06:00
Shangtong Zhang a4d41d522c Update bullet 2018-05-18 21:06:58 -06:00
Shangtong Zhang 938b7ce436 Merge branch 'ddpg-pixel' 2018-05-18 17:13:24 -06:00
Shangtong Zhang e19e49e9ba Automatic reset 2018-05-18 16:12:28 -06:00
Shangtong Zhang 6f029acf47 Tune params 2018-05-18 15:44:46 -06:00
Shangtong Zhang 20f34935e5 Update pixel ddpg params 2018-05-18 15:29:17 -06:00
Shangtong Zhang 93a5c3fc04 Minor update 2018-05-18 14:24:45 -06:00
Shangtong Zhang dcb363067e Update examples 2018-05-18 11:48:30 -06:00
Shangtong Zhang f6d0c4d260 Remove unused wrapper 2018-05-17 23:37:50 -06:00
Shangtong Zhang 84f7911bb4 Upgrade all actor critic methods 2018-05-17 23:37:17 -06:00
Shangtong Zhang aa07d467bd DDPG pixel 2018-05-17 17:32:01 -06:00
Shangtong Zhang 08038f9533 Upgrade DDPG 2018-05-17 17:17:12 -06:00
Shangtong Zhang ab1067d8a8 Fix frameskip 2018-05-17 14:00:03 -06:00
Shangtong Zhang 4699748624 Optimize replay buffer as baselines 2018-05-17 11:28:33 -06:00
Shangtong Zhang 85760246e1 Update logger 2018-05-11 16:35:20 -06:00
Shangtong Zhang aba27ae081 Add requirements.txt 2018-05-11 13:17:48 -06:00
Shangtong Zhang d33f85a2c5 Refine API 2018-05-11 09:28:42 -06:00
Shangtong Zhang 2ab14b83df Support docker 2018-05-10 22:30:49 -06:00
Shangtong Zhang 9d53c75bd8 Replace Mujoco with PyBullet 2018-05-10 22:30:31 -06:00
Shangtong Zhang 508ff8aea7 Rewrite process wrapper 2018-05-10 21:36:51 -06:00
Shangtong Zhang 9811e3b514 Remove object 2018-05-08 22:17:46 -06:00
Shangtong Zhang 53f44b09a3 Merge branch 'master' of https://github.com/ShangtongZhang/DeepRL 2018-05-08 15:39:19 -06:00
Shangtong Zhang 3929461f50 Use new style class 2018-05-08 15:38:48 -06:00
Shangtong Zhang 3c3a998c49 Update README 2018-05-07 22:03:11 -06:00
Shangtong Zhang 692a9098ff Minor update 2018-05-07 22:01:43 -06:00
Shangtong Zhang 06876d1269 Add Option-Critic 2018-05-07 22:01:21 -06:00
Shangtong Zhang fa294cd532 Fix import 2018-05-07 21:24:52 -06:00
Shangtong Zhang fb95820803 Package the project 2018-05-07 16:49:01 -06:00
Shangtong Zhang 8e11d762e7 Support python 2.7 2018-05-07 12:17:16 -06:00
Shangtong Zhang a6bcf9e5a7 Average over unaligned runs 2018-05-07 11:17:46 -06:00
Shangtong Zhang fb2dda9521 Parse args and upgrade mkdir 2018-05-01 21:41:55 -06:00
Shangtong Zhang 23279a4082 Merge branch 'master' of https://github.com/ShangtongZhang/DeepRL 2018-05-01 13:10:01 -06:00
Shangtong Zhang 205fc1d07c Update utils 2018-05-01 13:08:34 -06:00
Shangtong Zhang c6b694ac86 Use uuid4 2018-04-30 19:33:57 -06:00
Shangtong Zhang 94adf0e9a1 Set default gate to ReLU 2018-04-29 10:56:48 -06:00
Shangtong Zhang bc89bb2eed Init FCBody 2018-04-27 16:31:33 -06:00
Shangtong Zhang 6f3c26a5a8 Refactor fully-connected body 2018-04-27 16:29:22 -06:00