Commit Graph

1491 Commits

Author SHA1 Message Date
John Schulman
990c985717 Get rid of unnecessary indirection in Monitor 2017-02-11 17:40:51 -08:00
John Schulman
622429c39a improve toy_text printing 2017-02-11 17:39:57 -08:00
Tom Brown
4647fb1313 Mark current version as 0.7.4-dev 2017-02-09 11:15:18 -08:00
Tom Brown
a4dea774f2 Mark current version as 0.7.4-alpha 2017-02-09 11:13:39 -08:00
Tom Brown
35929e9669 Add zapier hooks to travis (#493) 2017-02-08 10:33:16 -08:00
John Schulman
385a85fd0c add MultiBinary space, which is a vector of {0,1} 2017-02-04 23:36:32 -08:00
John Schulman
9f23363e17 minor, spelling 2017-02-04 23:35:38 -08:00
Tom Brown
443d509df7 Add __version__ (#490) 2017-02-03 16:51:00 -08:00
Tom Brown
2d41d698c2 Unwrap env to access viewer 2017-02-01 15:54:41 -08:00
Tom Brown
d337f4e571 TimeLimit refactor with Monitor Simplification (#482)
* fix double reset, as suggested by @jietang

* better floors and ceilings

* add convenience methods to monitor

* add wrappers to gym namespace

* allow playing Atari games, with potentially more coming in the future

* simplify example in docs

* Move play out of the Env

* fix tests

* no more deprecation warnings

* remove env.monitor

* monitor simplification

* monitor simplifications

* monitor related fixes

* a few changes suggested by linter

* timestep_limit fixes

* keep track of gym env variables for future compatibility

* timestep_limit => max_episode_timesteps

* don't apply TimeLimit wrapper in make for VNC envs

* Respect old timestep_limit argument

* Pass max_episode_seconds through registration

* Don't include deprecation warnings yet
2017-02-01 13:10:59 -08:00
Tom Brown
c17ac6cc55 Bump version 2017-01-31 19:44:00 -08:00
catherio
ce441cd0f8 Logging message formatting typo 2017-01-31 13:58:30 -08:00
Szymon Sidor
f13b8cc699 fix benchmarks max_timesteps to be ints 2017-01-31 13:35:54 -08:00
Szymon Sidor
da2dbb9376 added exploration benchmark to Atari 2017-01-31 13:19:44 -08:00
Alejandro Dubrovsky
26556f99fe Fix multi-discrete sampling to be a proper uniform distribution 2017-01-30 16:14:17 -08:00
catherio
cde3b5e63b Bump version 2017-01-13 14:32:10 -08:00
catherio
061ca333d7 Expose ale lives in info dict 2017-01-13 14:09:56 -08:00
catherio
f412225f9e Revert "atari envs where life loss triggers new episode"
We intended to use this functionality in Universe, but have
decided to keep it within the Universe repo, and expose
ALE lives in the info dict instead (see next commit)
2017-01-13 14:09:20 -08:00
damodei
d626493705 atari envs where life loss triggers new episode 2017-01-12 23:19:22 -08:00
Tambet Matiisen
1183e8f6b5 Fixed timesteps in MinecraftHard-v0 benchmark. 2017-01-11 14:09:18 -08:00
vadimq
522c2c5322 Fix Taxi (upgrade -> v2) (#345)
* Disallow Taxi start location = goal location

* Bump version of Taxi env

* Update Taxi observations in the rollout

* Update Taxi reward threshold
2017-01-09 11:10:31 -08:00
John Schulman
518f4b7c4b change auto-configure behavior to unbreak universe 2017-01-04 17:29:32 -08:00
Szymon Sidor
a4eb1ac9d4 add ability to submit tags to benchmark_run using gym.upload 2017-01-04 10:41:45 +01:00
John Schulman
1a06fe0d9b get rid of auto reset 2017-01-03 23:51:25 -08:00
John Schulman
a67653ccb1 cartpole fix (#457)
minor fixes related to configure() and cartpole
2017-01-03 23:19:14 -08:00
John Schulman
c6f337e6b6 restore timestep limit in mujoco envs 2017-01-03 19:37:47 -08:00
futurely
bb9bbce80f Fix deprecated Monitor API for the cem example (#453) 2017-01-01 22:03:06 -08:00
Trevor Blackwell
ae8956921b Refactor json_encode_np 2016-12-30 12:06:07 -08:00
Trevor Blackwell
1059ccdf0a Handle numpy types in monitor
Fixes #452
2016-12-30 01:26:13 -08:00
Tom Brown
4cec14144b Add setter for backwards compat (#446)
* Add setter for backwards compat

* Fix ordering and preserver backwards compat
2016-12-28 15:38:55 -08:00
Tom Brown
133f1fdfdf Add timestep_limit property to EnvSpec to preserver backwards compat 2016-12-28 15:16:42 -08:00
Tom Brown
5ab1bab01c Bump version 2016-12-28 13:28:02 -08:00
Tom Brown
2d0b4e59e8 Improve error for removed decorator syntax 2016-12-28 13:28:02 -08:00
Tom Brown
109b7fb020 Update changelog 2016-12-28 13:24:43 -08:00
Greg Brockman
aae0c4c6f6 Tweak changelog 2016-12-28 13:18:01 -08:00
Tom Brown
69c1cda01c Downgrade DoubleWrap error to warning 2016-12-28 13:08:47 -08:00
Tom Brown
85da4b7877 Improve error message 2016-12-28 13:08:47 -08:00
Jie Tang
a81a631b2e Add error when wrapping after configure 2016-12-28 13:08:47 -08:00
Tom Brown
c3c703c6d8 Add configured tests 2016-12-28 13:08:47 -08:00
Jie Tang
10f7e4ffb0 Go from decorator to constructor syntax for Monitor; add note to What's New 2016-12-28 13:08:47 -08:00
Jie Tang
cf77e19c84 Update comment 2016-12-28 13:08:47 -08:00
Tom Brown
68c66b9a00 Kill timestep_limit 2016-12-28 13:08:47 -08:00
Tom Brown
2890611741 timestep_limit => tags 2016-12-28 13:08:47 -08:00
Jie Tang
0d02a34de2 Remove timestep enforcement from monitor, adjust TimeLimit semantics to match existing monitor, fix test for env reuse 2016-12-28 13:08:47 -08:00
Jie Tang
a4637dc1ee remove typo 2016-12-28 13:08:47 -08:00
Jie Tang
557c25ea5f Add TimeLimit wrapper to gym, make default monitor wrapper use TimeLimit, add test for TimeLimit reset behavior 2016-12-28 13:08:47 -08:00
Ben Giles
d481151470 Fix CEM example for Monitor wrapper change #434 2016-12-28 00:32:41 -08:00
bstriner
5afa0b4e2c Skip setsid if attribute not present 2016-12-28 00:31:31 -08:00
Tom Brown
3ad93f62fd Improve deprecation message 2016-12-27 20:02:33 -08:00
Sung Kim
63c4fb965e Fixed RIGHT/LEFT/DOWN/UP for Frozen Lake 2016-12-25 01:06:36 -08:00