Zach Dwiel
499e78596a
Merge pull request #38 from NervanaSystems/nec_lstm
...
update nec and value optimization agents to work with recurrent middleware
2018-01-08 14:01:34 -05:00
Justin
29857412b3
Add force flag to library symbolic link
...
Link step fails if continuing installation after interruption, requiring manual deletion of the link. Adding a force flag overrides the existing symbolic link from attempted installation in the newly created virtual environment.
2018-01-08 20:36:54 +02:00
Zach Dwiel
6c79a442f2
update nec and value optimization agents to work with recurrent middleware
2018-01-05 20:16:51 -05:00
Itai Caspi
645d9d47a9
Adding bibtex to the README
2018-01-03 21:11:57 +02:00
Itai Caspi
93a54c7e8e
Added a link to the 2nd blog post
2017-12-20 17:18:49 +02:00
Itai Caspi
9e59d1960e
bug fix for dumping gifs from doom
2017-12-20 13:10:34 +02:00
Zach Dwiel
37e317682b
allow missing carla environment and missing matplotlib package
2017-12-20 11:47:14 +02:00
Itai Caspi
125c7ee38d
Release 0.9
...
Main changes are detailed below:
New features -
* CARLA 0.7 simulator integration
* Human control of the game play
* Recording of human game play and storing / loading the replay buffer
* Behavioral cloning agent and presets
* Golden tests for several presets
* Selecting between deep / shallow image embedders
* Rendering through pygame (with some boost in performance)
API changes -
* Improved environment wrapper API
* Added an evaluate flag to allow convenient evaluation of existing checkpoints
* Improve frameskip definition in Gym
Bug fixes -
* Fixed loading of checkpoints for agents with more than one network
* Fixed the N Step Q learning agent python3 compatibility
2017-12-19 19:27:16 +02:00
Itai Caspi
11faf19649
QR-DQN bug fix and imporvements ( #30 )
...
* bug fix - QR-DQN using error instead of abs-error in the quantile huber loss
* improvement - QR-DQN sorting the quantile only once instead of batch_size times
* new feature - adding the Breakout QRDQN preset (verified to achieve good results)
2017-11-29 14:01:59 +02:00
Zach Dwiel
7bdba396d2
Update add_env.md
2017-11-14 17:57:55 +02:00
Zach Dwiel
9ae2905a76
clean up input embeddings setup
2017-11-14 17:39:18 +02:00
Itai Caspi
1ff0da2165
bug fix - fixed an issue with gifs dumping and bumped up Pillow version to 4.3.0
2017-11-13 12:22:42 +02:00
Miguel Morales
acd2b78a9e
Update README.md
...
Fix algorithms list to be consistent with "<full name> (<acronym>)"
2017-11-12 16:00:00 +02:00
Itai Caspi
8d9ee4ea2b
bug fix - fixed C51 presets hyperparameters
2017-11-10 13:22:42 +02:00
galleibo-intel
3c330768f0
Fix for NEC not saving the DND when saving a model
2017-11-09 19:13:23 +02:00
Itai Caspi
f5d645d8a6
resize training curves images
2017-11-09 09:13:12 +02:00
Itai Caspi
8ee9e46083
fixing some typos in the benchmarks README
2017-11-09 08:58:52 +02:00
Itai Caspi
c798be7bfb
added training curves for some of the presets
2017-11-09 08:54:34 +02:00
cxx
84e536d371
Fix std calculation using unbiased estimation in sharing stat mode.
2017-11-07 20:19:54 +02:00
galleibo-intel
f47b8092af
fix for intel optimized tensorflow on distributed runs + adding coach_env to .gitignore
2017-11-06 19:41:32 +02:00
Itai Caspi
b40259c61a
bug fix - remove import warning when everything was imported successfully + changed global step api to match TF 1.4
2017-11-06 17:28:13 +02:00
Itai Caspi
fd103a7b69
updated the algorithms diagram with QR-DQN
2017-11-01 15:24:54 +02:00
Itai Caspi
a8bce9828c
new feature - implementation of Quantile Regression DQN ( https://arxiv.org/pdf/1710.10044v1.pdf )
...
API change - Distributional DQN renamed to Categorical DQN
2017-11-01 15:09:07 +02:00
Itai Caspi
1ad6262307
bug fix - correcting the evaluation exploration control parameter logging
2017-10-31 13:50:40 +02:00
Itai Caspi
e38611b9eb
bug fix - updating Doom_Health_DFP and Breakout_DQN presets
2017-10-31 10:54:14 +02:00
Itai Caspi
913ab75e8a
bug fix - preventing crashes when the probability of one of the actions is 0 in the policy head
2017-10-31 10:51:48 +02:00
Itai Caspi
1918f16079
imporved API for getting / setting variables within the graph
2017-10-31 10:51:48 +02:00
cxx
e33b0e8534
Fix preset mistakes.
2017-10-26 12:37:32 +03:00
cxx
f43c951c2d
Unify base class using new-style (object).
2017-10-26 12:33:09 +03:00
Itai Caspi
39cf78074c
preventing the evaluation agent from getting stuck in bad policies by updating from the global network during episodes
2017-10-25 10:28:45 +03:00
Itai Caspi
d3c6860421
removing some Gym wrapper code that was required for OpenAI Universe in the past
2017-10-25 10:01:10 +03:00
Gal Leibovich
eb0b57d7fa
Updating PPO references per issue #11
2017-10-24 16:57:44 +03:00
Itai Caspi
43bc359166
updated atari presets with v4 environment ids
2017-10-23 14:14:09 +03:00
Gal Leibovich
1a09b7cec3
changing python to python3 everywhere to make the supported version of python explicit
2017-10-23 13:07:54 +03:00
Gal Leibovich
16604b11df
Update install.sh for missing AEON dependencies
...
AEON (nervana data loader) is being installed as part of the neon installation
2017-10-23 11:50:41 +03:00
Itai Caspi
a1656c2ae6
fixed docs color for mobile
2017-10-23 11:46:27 +03:00
Gal Leibovich
7c5d0ed01b
Update install.sh to install virtualenv when installing neon
...
this is required for a case the user asks the user *not* to install Coach in a virtualenv (and then the installer does not install virtualenv, which is required for neon)
2017-10-23 10:51:48 +03:00
Zach Dwiel
aceddfadc9
Merge pull request #3 from cclauss/patch-1
...
Simplify w/ dict.get() default value, ternary if
2017-10-22 20:46:36 -04:00
Zach Dwiel
6fec9f972f
Merge pull request #7 from zac-hopkinson/master
...
Fixes installer issues #5 and #6
2017-10-22 20:42:57 -04:00
Gal Novik
6009b73eb6
fixed some documentation typos
2017-10-22 22:21:45 +03:00
Zac Hopkinson
2a48d5fb66
Fixes installer issues #5 and #6
2017-10-22 14:54:45 -04:00
Gal Leibovich
2a3a6f4a68
small fix in neon installation
2017-10-22 16:43:02 +03:00
Gal Leibovich
cc9580a949
updated docs with links to github + a few more words on Dashboard functionality
2017-10-22 16:33:49 +03:00
Itai Caspi
1e47257178
fixed gym installation (failed due to packages removal from apt-get repo)
2017-10-22 15:05:25 +03:00
Itai Caspi
d976609127
fixed a typo in the installation script
2017-10-22 12:24:22 +03:00
Itai Caspi
497130098c
renamed NGraph to neon in the installation script
2017-10-22 11:57:02 +03:00
Gal Leibovich
c3501653f7
Update README.md
2017-10-22 09:09:37 +03:00
cclauss
10c139a28c
Update utils.py
2017-10-22 07:42:33 +02:00
cclauss
6e9275edc3
Simplify w/ dict.get() default value, ternary if
2017-10-22 07:41:07 +02:00
Itai Caspi
aacd9b5db8
Fixed link to MMC in the README
2017-10-21 20:26:45 +03:00
