Sign upLogin

kengz/SLM-Lab

View on GitHub
Star

All Builds

Merge remote-tracking branch 'origin/master' into a2c-vec-act #1708
8209263kengz authored 35aa7899
complete00:00:30
Merge pull request #300 from kengz/net-out-activ #1707
8209263Wah Loon Keng authored f8567e3a
complete00:00:17
fix hydra #1706
8209263kengz authored 2d27f5ba
complete00:00:24
Merge pull request #296 from kengz/remove-inplace #1705
8209263Wah Loon Keng authored 92af65bc
complete00:00:23
add out_layer_activation to networks #1704
8209263kengz authored d4f36c2f
complete00:04:38
Check orthogonal init again after changes #1703
5512945lgraesser authored b776e476
complete00:00:37
Change default init back to none, guard against gain for xavier_uniform_ #1702
5512945lgraesser authored 5b142bad
complete00:00:34
Merge pull request #296 from kengz/remove-inplace #1701
8209263Wah Loon Keng authored 92af65bc
complete00:00:18
Merge branch 'master' into dqn-atari-exps #1700
8209263Wah Loon Keng authored a69a5991
complete00:00:18
Merge pull request #290 from kengz/fix-ddqn #1699
8ad2cd696016bcbe4ca9ce3870d96c43Laura Graesser authored 590cd427
complete00:00:16
change spec names #1698
5512945lgraesser authored 65e6f336
complete00:00:20
A2C pong specs #1697
5512945lgraesser authored ba3cfb8a
complete00:00:28
Merge branch 'fix-ppo-keng' of https://github.com/kengz/SLM-Lab into fix-ppo-keng #1696
5512945lgraesser authored 8f72428c
complete00:00:19
Merge branch 'fix-ppo-keng' of https://github.com/kengz/SLM-Lab into fix-ppo-keng #1695
8209263kengz authored 85cc27d3
complete00:00:24
Cleanup #1694
5512945lgraesser authored a3003123
complete00:00:15
move log loop out #1693
8209263kengz authored 88a73d99
complete00:00:17
fix control and loggin #1692
8209263kengz authored 73ed9fc2
complete00:00:20
Merge pull request #296 from kengz/remove-inplace #1691
8209263Wah Loon Keng authored 92af65bc
complete00:00:21
hack API to make venv control work #1690
8209263kengz authored 8367bf53
complete00:00:18
Merge branch 'master' into fix-ppo-keng #1689
8209263Wah Loon Keng authored 46095f1f
complete00:00:23
Merge pull request #296 from kengz/remove-inplace #1688
8209263Wah Loon Keng authored 92af65bc
complete00:00:22
remove untested init #1687
8209263kengz authored 7438b507
complete00:00:21
guard kaiming uniform init #1686
8209263kengz authored 87bb2ef0
complete00:00:22
eval demo less frequently, lower clip grad #1685
8209263kengz authored cffb7449
complete00:00:21
update demo dqn spec #1684
8209263kengz authored 1c9596ab
complete00:00:18
fix weight init #1683
8209263kengz authored db922be3
complete00:00:21
sq 32 #1682
5512945lgraesser authored 67489d9b
complete00:00:17
more steps #1681
5512945lgraesser authored c3f208da
complete00:00:18
more env steps for stability (?) #1680
5512945lgraesser authored 71e4efbe
complete00:00:20
a2c seaquest #1679
5512945lgraesser authored 6a64246b
complete00:00:22
gae pong spec #1678
5512945lgraesser authored ff0e0313
complete00:00:20
use slow but comparable nstep calc #1677
8209263kengz authored 281d1938
complete00:00:23
Merge branch 'fix-ppo-keng' of https://github.com/kengz/SLM-Lab into fix-ppo-keng #1676
5512945lgraesser authored 4c842516
complete00:00:18
Merge remote-tracking branch 'origin/master' into fix-ppo-keng #1675
8209263kengz authored 252b1ab8
complete00:00:20
Change a2c_pong to CategoricalCustom #1674
5512945lgraesser authored aa9dae0c
complete00:00:19
Merge pull request #295 from kengz/refactor #1673
8209263Wah Loon Keng authored 10fbb7fd
complete00:00:16
relax grad norm test #1672
8209263kengz authored ca50eb6b
complete00:00:17
remove retired code #1671
8209263kengz authored ca43d0e6
complete00:00:24
Merge pull request #290 from kengz/fix-ddqn #1670
8ad2cd696016bcbe4ca9ce3870d96c43Laura Graesser authored 590cd427
complete00:00:24
Noisy logits #1669
5512945lgraesser authored 7d09e1b2
complete00:00:29
fix init_layers nonlinearity names #1668
8209263kengz authored 6ff72079
complete00:00:16
fix polyak typo #1667
8209263kengz authored 8b3c8651
complete00:00:23
fix style #1666
8209263kengz authored dbc7cfd1
complete00:00:22
use get_nn_name in RNN cell #1665
8209263kengz authored 4d950189
complete00:00:26
np import #1664
5512945lgraesser authored 10234de0
complete00:00:29
Merge pull request #290 from kengz/fix-ddqn #1663
8ad2cd696016bcbe4ca9ce3870d96c43Laura Graesser authored 590cd427
complete00:00:23
Openai net init #1662
5512945lgraesser authored 7ba764f1
complete00:00:33
Merge pull request #290 from kengz/fix-ddqn #1661
8ad2cd696016bcbe4ca9ce3870d96c43Laura Graesser authored 590cd427
complete00:00:24
update a2c pong spec #1660
5512945lgraesser authored b97e9f56
complete00:00:24
Dont cross episode boundaries in batch replay #1659
5512945lgraesser authored 41b1eec7
complete00:00:21