Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Feb 20, 2018

• * added citation

  jingweiz

  committed
  20b3b73

  Copy full SHA for 20b3b73

• Merge branch 'master' of https://github.com/jingweiz/pytorch-rl

  jingweiz

  committed
  bddeed4

  Copy full SHA for bddeed4

Commits on Nov 21, 2017

• Merge pull request #12 from jingweiz/revert-10-master

  Show description for 5b45311
  jingweiz

  authored
  5b45311

  Copy full SHA for 5b45311

• Revert "Added opensim-rl environment and a sample configuration and options for a continuous DQN agent to learn in that environment"

  jingweiz

  authored
  00c6c90

  Copy full SHA for 00c6c90

Commits on Nov 20, 2017

• Merge pull request #10 from praveen-palanisamy/master

  Show description for a990c7d
  jingweiz

  authored
  a990c7d

  Copy full SHA for a990c7d

Commits on Oct 22, 2017

• Updated README to add the optional dependency on opensim-rl environment

  praveen-palanisamy

  committed
  cd3b9cf

  Copy full SHA for cd3b9cf

• Modified the DQN agent so that it can produce a list of actions (useful for continuous action spaces). Tested with the opensim-rl environment which has a continous action space

  praveen-palanisamy

  committed
  e311708

  Copy full SHA for e311708

• Added factory configurations and options to train a continuous DQN agent in the opensim-rl envrionment

  praveen-palanisamy

  committed
  7391551

  Copy full SHA for 7391551

• Added opensim-rl envrionment to the list of available environments.In this environment, the agent is tasked to learn to run while avoiding obstacles on the ground

  praveen-palanisamy

  committed
  fcdab5b

  Copy full SHA for fcdab5b

Commits on Aug 27, 2017

• @ aisgpu8 17082701

  Show description for d9717ef
  jingweiz

  committed
  d9717ef

  Copy full SHA for d9717ef

• @ aisgpu7 17082700

  Show description for 17bdccc
  jingweiz

  committed
  17bdccc

  Copy full SHA for 17bdccc

• * fixes

  jingweiz

  committed
  67e1bee

  Copy full SHA for 67e1bee

Commits on Aug 24, 2017

• * fixed preprocess for off-policy rollout

  jingweiz

  committed
  7c03adf

  Copy full SHA for 7c03adf

Commits on Aug 23, 2017

• * fixes

  jingweiz

  committed
  8f953c0

  Copy full SHA for 8f953c0

• * small mods

  jingweiz

  committed
  9cea36c

  Copy full SHA for 9cea36c

• * updated readme

  jingweiz

  committed
  c33e04f

  Copy full SHA for c33e04f

• * finished up acer discrete version

  jingweiz

  committed
  5681952

  Copy full SHA for 5681952

• * now can choose to log at train_step or frame_step

  jingweiz

  committed
  dd51b10

  Copy full SHA for dd51b10

• * added some loggings

  jingweiz

  committed
  f62d04d

  Copy full SHA for f62d04d

• * updated readme

  jingweiz

  committed
  4d3533f

  Copy full SHA for 4d3533f

• * cleanups

  jingweiz

  committed
  1a92f5d

  Copy full SHA for 1a92f5d

• * clean up backward and rm unncessary cat by directly outputting unsplitted policy_vb from _off_policy_rollout

  jingweiz

  committed
  6ec6ea0

  Copy full SHA for 6ec6ea0

• * cleared up backward procedures

  jingweiz

  committed
  254f2b6

  Copy full SHA for 254f2b6

Commits on Aug 21, 2017

• * comments

  jingweiz

  committed
  b9b87d8

  Copy full SHA for b9b87d8

• * rm old comment

  jingweiz

  committed
  907107b

  Copy full SHA for 907107b

• * entropy loss & value loss

  jingweiz

  committed
  c6e4edc

  Copy full SHA for c6e4edc

• * w/ & w/p trpo runnable

  jingweiz

  committed
  7ec9b42

  Copy full SHA for 7ec9b42

• * small mods

  jingweiz

  committed
  b410023

  Copy full SHA for b410023

• * 1st order trpo to only return z, backward is done outside on the whole rollout instead of individually

  jingweiz

  committed
  f5ee762

  Copy full SHA for f5ee762

Commits on Aug 20, 2017

• * tidy up & added comments

  jingweiz

  committed
  ff10e13

  Copy full SHA for ff10e13

• * bias correction & 1st order trpo

  jingweiz

  committed
  ebc040f

  Copy full SHA for ebc040f

Commits on Aug 17, 2017

• * adjust lr

  Show description for 4e6abab
  jingweiz

  committed
  4e6abab

  Copy full SHA for 4e6abab

• * tidy up coding style

  jingweiz

  committed
  d78f831

  Copy full SHA for d78f831

Commits on Aug 16, 2017

• * draft of _forward passes

  jingweiz

  committed
  4a18221

  Copy full SHA for 4a18221

Commits on Aug 15, 2017

• * added some todos

  jingweiz

  committed
  09aab4f

  Copy full SHA for 09aab4f

Pagination

Previous

Next