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
55ae92619a
Update README
2017-10-28 11:08:25 -06:00
Shangtong Zhang
9615195f5e
Upgrade to PyTorch v0.2.0
2017-10-28 11:06:20 -06:00
Shangtong Zhang
a59f2d2911
Benchmark DDPG with Roboschool
2017-10-27 23:09:24 -06:00
Shangtong Zhang
af360bb9ea
Support roboschool
2017-10-27 14:23:54 -06:00
Shangtong Zhang
e832e2e89a
Update README
2017-10-26 23:06:20 -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
c04b2a395f
Update README
2017-10-10 23:02:28 -06:00
Shangtong Zhang
51f125c01e
Improve normalizer for scalar
2017-10-07 21:23:02 -06:00
Shangtong Zhang
fae9a85f31
Major update
2017-10-06 22:10:44 -06:00
Shangtong Zhang
8189a5d136
Update README
2017-10-06 11:48:58 -06:00
Shangtong Zhang
c5cbc10f94
Refactor DDPG
2017-10-06 11:16:21 -06:00
Shangtong Zhang
8b9fd8d24f
Code cleanup
2017-10-05 21:42:27 -06:00
Shangtong Zhang
dd2443c1c3
DPPO
2017-10-05 21:40:42 -06:00
Shangtong Zhang
e58938e3fe
Support async PPO, which doesn't work
2017-10-04 22:33:26 -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
2b30c6e999
Update README
2017-08-03 21:02:03 -06:00
Shangtong Zhang
44e8a782af
Refactor DQN
2017-08-02 14:55:04 -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
473304d9a7
Update README
2017-07-30 13:36:15 -06:00
Shangtong Zhang
bbfcbf76f4
Major reversion
2017-07-30 13:24:04 -06:00
Shangtong Zhang
ce504e2d0f
Major refactor
2017-07-26 18:18:49 -06:00
Shangtong Zhang
fb71f51ea7
Refactor file structure
2017-07-25 21:00:13 -06:00
Shangtong Zhang
c79cdb18ea
Refactor async agent
2017-07-25 12:06:54 -06:00
Shangtong Zhang
8b700774d2
Update logger
2017-07-22 19:03:07 -06:00
Shangtong Zhang
2e6f49db85
Fix failed softmax layer test
2017-07-22 10:07:30 -06:00
Shangtong Zhang
709cde1d4b
Introduce tensorboard
2017-07-22 09:28:02 -06:00
Shangtong Zhang
a145a01cb6
Minor refactor
2017-07-20 20:47:51 -06:00
Shangtong Zhang
7b7d37694e
Update README
2017-07-19 23:19:11 -06:00
Shangtong Zhang
5447d64caa
DDPG
2017-07-19 23:11:50 -06:00
Shangtong Zhang
0b88546091
Fix bug
2017-07-19 10:29:29 -06:00
Shangtong Zhang
f88b565f6b
Update figures for A3C
2017-06-19 09:24:16 -06:00
Shangtong Zhang
b181f109f5
Major refactor
2017-06-18 20:51:27 -06:00
Shangtong Zhang
8ea108e4bb
Finalize async methods
2017-06-10 21:08:40 -06:00
Shangtong Zhang
9c73abfbff
Major refactor to support LSTM layer
2017-06-08 18:36:04 -06:00
Shangtong Zhang
af17bbbf94
Tune parameters
2017-06-05 15:23:41 -06:00
Shangtong Zhang
a9d4a0fa72
New network for value based async methods
2017-06-04 19:21:04 -06:00
Shangtong Zhang
b8b7602b35
Update README
2017-06-04 14:47:22 -06:00
Shangtong Zhang
1840438561
Update README
2017-06-04 14:43:36 -06:00
Shangtong Zhang
20871f918f
Update figures
2017-06-04 14:19:37 -06:00
Shangtong Zhang
1aebec066c
Update Readme
2017-06-03 15:07:53 -06:00