Jie Tang
a8e4734cbc
Fix evaluation scoring bug (numpy casts generator to a single-element array containing the generator, which is a truthy object)
2016-10-20 19:39:20 -07:00
Greg Brockman
88f94587a2
Update benchmark spec ( #385 )
...
* Update benchmark spec
* Update format of benchmark again
* Add support for max_seconds to benchmark
* Bump version
2016-10-20 17:25:29 -07:00
John Schulman
21f8fe57be
fix to previous commit
2016-10-20 11:41:23 -07:00
John Schulman
3546f06bc3
prevents race condition in multiprocessing setting
2016-10-20 11:40:20 -07:00
John Schulman
50be1426be
fix number of seeds in mujoco benchmark
2016-10-20 11:39:03 -07:00
Jie Tang
a8aa301202
Bump version
2016-10-20 01:34:33 -07:00
Trevor Blackwell
768f865410
Moved to gym-recording repo
2016-10-19 16:29:22 -07:00
Greg Brockman
517c2e0524
Tweak error message
2016-10-18 20:29:38 -07:00
Jie Tang
db4f977bcd
Treat each directory as a single trial
2016-10-18 20:26:08 -07:00
Jie Tang
2dc3e56ac2
Don't check exact seeds for now
2016-10-18 12:05:01 -07:00
Jie Tang
8aa22a13b9
Bump version
2016-10-18 11:40:12 -07:00
jietang
cc0690cf1a
Add support for configuring monitor output name ( #379 )
...
* configurable monitor filename
* removing debug logging
* Add documentation for uid argument to monitor.start
2016-10-18 11:37:29 -07:00
Jie Tang
d35c2f0480
minor tweak to benchmark runner helper
2016-10-17 17:56:14 -07:00
Jie Tang
6e0f83e9d5
Validate multiple seeds for benchmark
2016-10-17 17:56:14 -07:00
Jie Tang
8941313121
Make random benchmark runner script respect seeds
2016-10-17 17:56:14 -07:00
Greg Brockman
65b3454a77
Bump version
2016-10-17 14:46:35 -07:00
Greg Brockman
a86b4ba098
bump version
2016-10-15 00:23:43 -07:00
damodei
5afbb71929
add reverse action filter for demos
2016-10-14 22:07:47 -07:00
Jie Tang
1ef455dab9
Restore imports
2016-10-14 00:10:02 -07:00
Jie Tang
a771a7306e
Remove benchmark without envs
2016-10-13 23:39:32 -07:00
Jie Tang
cca6ef5914
Add benchmark names, add tinkerbell benchmarks
2016-10-11 20:58:11 -07:00
Jie Tang
953d56cb18
Example benchmark runner script
2016-10-09 20:49:53 -07:00
Jie Tang
422882913b
Separate benchmark and evaluation control flow in upload(), add client side validation of benchmark contents
2016-10-09 20:49:49 -07:00
Jie Tang
2933a95ec5
Adding a simple classic control benchmark
2016-10-04 23:03:32 -07:00
Trevor Blackwell
838f0ad1e6
Improve doc
2016-10-03 16:31:39 -07:00
Greg Brockman
2541b0683a
Add back extra argument
2016-10-01 21:59:08 -07:00
Greg Brockman
3568b760b5
Bump version
2016-10-01 19:37:32 -07:00
Greg Brockman
483c7805c0
Improve logging configuration
2016-10-01 19:35:41 -07:00
Greg Brockman
9a5c0fc63e
Add support for extra loggers
2016-10-01 19:24:50 -07:00
Greg Brockman
0c4cc9c2d9
Bump version
2016-09-30 14:49:27 -07:00
Greg Brockman
66bcbcdab6
Move repeat_action_probability before loading the ROM ( #362 )
...
Closes #349
2016-09-29 02:49:48 -07:00
Greg Brockman
d831826ec2
Downgrade to just setting the gym logger
2016-09-28 22:58:57 -07:00
catherio
aa3fad8e00
Refactor
2016-09-28 10:24:44 -07:00
catherio
3cd5b4038b
No empty dicts or lists as default values
...
Apparently python does not play nicely with the empty dict (or the
empty list for that matter) as a default value:
http://stackoverflow.com/questions/5712904/empty-dictionary-as-default-value-for-keyword-argument-in-python-function-dicti
2016-09-27 23:13:11 -07:00
catherio
0900e7d2fd
Tags are now a dict; make sure to initialize as a dict, not a list
2016-09-27 22:59:39 -07:00
Greg Brockman
8a81a20827
Bring back old Atari-v0 variants
2016-09-24 15:55:56 -07:00
Greg Brockman
45038020ae
Assign floor for any missing episodes
2016-09-23 02:08:11 -07:00
Nestor Gonzalez
312c8b439c
added link to gym-gazebo ( #356 )
2016-09-23 01:59:41 -07:00
Greg Brockman
2b3f965faa
Fix scoring when fewer episodes are provided
2016-09-23 01:47:42 -07:00
Greg Brockman
934b2acbb7
Add benchmark support ( #338 )
...
* Warn if seed doesn't return a list
* Add preliminary BenchmarkRun support
* Add experimental benchmark registration
* Flesh out interface
* Add preliminary BenchmarkRun support
* Warn if seed doesn't return a list
* Add experimental benchmark registration
* Flesh out interface
* Make benchmarkrun upload recursive
* Add evaluation episodes
* Add benchmark scoring
* Tweak reward locations
* Tweak scoring
* Clear default metadata in Wrapper
* Improve scoring
* Expose registry; fix test
* Add initial_reset_timestamp
* Add back algorithm; fix tests
2016-09-23 01:04:26 -07:00
Paul Hendricks
0c8c055fca
added link to gym-inventory, a repository of inventory control environments ( #355 )
2016-09-22 21:26:41 -07:00
Greg Brockman
618eecc41e
Bump version
2016-09-21 14:56:19 -07:00
Greg Brockman
ada730a5b1
Switch the Gym automated logger setup to configure the root logger rather than just the 'gym' logger
...
Python doesn't make it easy for libraries to take responsibility for
logging configuration (which we do to make simple usage much easier),
and as we see more Gym plugins, we want their loggers to have an
appropriate log level too. So we may as well configure the root logger
level.
2016-09-21 14:55:04 -07:00
Greg Brockman
46b47725a3
Bump version
2016-09-21 00:37:37 -07:00
Greg Brockman
2761741ad3
Disable ALE's underlying repeat_action_probability ( #354 )
...
* Disable ALE's underlying repeat_action_probability
* Add assertion for repeat_action_probability
* Properly encode for py3
2016-09-21 00:36:56 -07:00
Greg Brockman
3c95d3af8f
Clear default metadata in Wrapper
...
Closes #340
2016-09-09 10:13:53 -07:00
Greg Brockman
bdba063624
Warn if seed doesn't return a list
2016-09-09 10:13:41 -07:00
Trevor Blackwell
ddcdbcd067
Merge branch 'trace_recording'
2016-09-07 19:44:26 -07:00
vadimq
46b48a8a4a
Fix Taxi-v1 for Python 3 ( #327 )
...
* Fix Taxi for Python3
* Update Taxi observations in the rollout
2016-09-07 19:43:45 -07:00
Trevor Blackwell
17bde31080
Merge branch 'master' into trace_recording
2016-09-07 19:41:46 -07:00
