45 Commits

Include renames

Author	SHA1	Message	Date
Shangtong Zhang	45587b9cce	Update PPO	2018-04-08 09:14:35 -06:00
Shangtong Zhang	8dfe6ff7c8	Rewrite state/reward normalizer	2018-04-06 11:36:44 -06:00
Shangtong Zhang	4c6481d5be	Tune parameters for DDPG	2018-04-06 10:34:52 -06:00
Shangtong Zhang	549e4af3b8	Rewrite PPO	2018-04-05 22:32:34 -06:00
Shangtong Zhang	8ad31c79b8	Update DDPG	2018-04-04 21:54:26 -06:00
Shangtong Zhang	0b57fdbc70	Rewrite discrete networks	2018-04-04 16:19:40 -06:00
Shangtong Zhang	63f8027199	Use openai atari wrapper	2018-04-03 20:20:18 -06:00
Shangtong Zhang	e7e94d715b	Add keepdim	2018-03-31 21:46:16 -06:00
Shangtong Zhang	0eaba783e8	Minor update	2018-03-24 23:21:58 -06:00
Shangtong Zhang	6863fc9c14	QR DQN Atari	2018-03-21 10:55:59 -06:00
Shangtong Zhang	4fbf811188	Quantile Regression DQN cartpole	2018-03-21 10:39:00 -06:00
Shangtong Zhang	f61908ba2a	Categorical DQN for Atari	2018-03-12 21:03:38 -06:00
Shangtong Zhang	5b4ecd882e	Fix a bug of Categorical DQN	2018-03-12 20:37:12 -06:00
Shangtong Zhang	3e47451ef6	Categorical DQN cart pole	2018-03-11 23:54:35 -06:00
Shangtong Zhang	6de8062702	Code cleanup	2018-02-12 14:27:40 -07:00
Shangtong Zhang	790c8f5cb9	Orthogonal init	2018-02-12 11:58:30 -07:00
Shangtong Zhang	b24593352d	Remove PIL	2018-02-09 16:34:30 -07:00
Shangtong Zhang	e91ec3be45	Specify a gpu for a network	2018-02-03 10:53:14 -07:00
Shangtong Zhang	fef04b6bf6	Deepmind actor-critic net	2018-02-03 09:51:52 -07:00
Shangtong Zhang	049e56e976	Minor update	2018-02-02 15:19:55 -07:00
Shangtong Zhang	ad9189e25a	Code cleanup	2018-01-31 21:00:19 -07:00
Shangtong Zhang	797f80d5ff	Code cleanup	2018-01-31 20:14:33 -07:00
Shangtong Zhang	ceaf83bca3	Support A2C	2018-01-31 16:19:33 -07:00
Shangtong Zhang	63c5d5ea54	Network cleanup	2018-01-19 11:29:05 -07:00
Shangtong Zhang	eb16eeb2a1	Update DDPG	2018-01-19 11:24:32 -07:00
Shangtong Zhang	69b7ea53cc	Minor update for DQN	2017-12-21 22:22:36 -07:00
Shangtong Zhang	13533ab2fe	Set hidden size	2017-12-18 20:09:11 -07:00
Shangtong Zhang	83a4270bb2	Upgrade to PyTorch v0.3	2017-12-11 23:41:16 -07:00
Mike Clark	deab7043fc	add eps to avoid NaN ... I tracked some NaN's I was getting down to here. It happens when std is a small number, then var is even smaller, and log_density=inf. There are some problems where the agent will learn to use small standard deviations because jittering movements have a high cost, this makes those more stable.	2017-11-04 16:24:54 +08:00
Mike Clark	212ff0bc70	moving epsilon outside softplus ... I think you intended this to be outside the softplus. The reason is that it should be applied just before the log to avoid `log(0)=inf`.e.g. - `log(softplus(-1000+1e-5))=log(0)=inf`. - `log(softplus(-1000)+1e-5)=log(1e-5)!=inf`. Also this fixes a NaN I had.	2017-11-01 10:52:23 +08:00
Shangtong Zhang	a59f2d2911	Benchmark DDPG with Roboschool	2017-10-27 23:09:24 -06:00
Shangtong Zhang	6682937b54	Support python3 and pytorch 0.2	2017-10-26 23:01:57 -06:00
Shangtong Zhang	1b6de16b4f	Fix a critical bug in DDPG	2017-10-15 10:05:33 -06:00
Shangtong Zhang	3ed6f2a9db	Log wall time	2017-10-14 16:38:51 -06:00
Shangtong Zhang	c5cbc10f94	Refactor DDPG	2017-10-06 11:16:21 -06:00
Shangtong Zhang	c773da7d08	Support shared stats for continuous A3C	2017-10-04 20:58:42 -06:00
Shangtong Zhang	3935dfc428	Unifying networks for continuous A3C and PPO	2017-10-04 11:03:02 -06:00
Shangtong Zhang	3adceb5284	Gaussian actor	2017-10-04 10:15:28 -06:00
Shangtong Zhang	27f54ed420	Single thread PPO	2017-10-04 09:33:35 -06:00
Shangtong Zhang	e60e9feecb	Refactor for HRA	2017-08-29 22:10:13 -06:00
Shangtong Zhang	d52182882a	Implementation of hybrid reward architecture	2017-08-29 21:52:26 -06:00
Shangtong Zhang	b161200f0f	Add BN layer	2017-08-02 14:35:51 -06:00
Shangtong Zhang	5116733f22	DDPG Pendulum	2017-08-01 10:52:14 -06:00
Shangtong Zhang	1be3b44999	Continuous A3C	2017-07-31 22:53:09 -06:00
Shangtong Zhang	ce504e2d0f	Major refactor	2017-07-26 18:18:49 -06:00