Commits

Breadcrumbs

History for

onmain

User selector

All users

Datepicker

All time

Commit History

Commits on Jul 27, 2021

• adds workaround for CSim 4_1 task crashes (mainly in tasks other than ReachTarget)

  deepdad

  committed
  6fb3fce

  Copy full SHA for 6fb3fce

• WipeDesk crashes, possibly due to using CSim 4.2

  deepdad

  committed
  062cf94

  Copy full SHA for 062cf94

• cleanup for submission

  deepdad

  committed
  c6b2cab

  Copy full SHA for c6b2cab

• style(dreamer-pytorch): cleaned up the code for submission (disables Atari/DMC)

  Show description for 96dfef7
  deepdad

  committed
  96dfef7

  Copy full SHA for 96dfef7

• fix(dense.py): Changes the Bernoulli distribution to ContinuousBernoulli after a test run

  Show description for 73236b4
  deepdad

  committed
  73236b4

  Copy full SHA for 73236b4

Commits on Jul 24, 2021

• really merges some clean up that was done in the imitation learning branch

  deepdad

  committed
  1e0288c

  Copy full SHA for 1e0288c

• changes action space print message

  deepdad

  committed
  a068771

  Copy full SHA for a068771

• imitation learning is starting to work, but it's still not tested, at least loading a prefilled model works, but it barely apermits changing parameters: that would require a state_dict wrapper that…

  Show description for 50d6998
  deepdad

  committed
  50d6998

  Copy full SHA for 50d6998

Commits on Jul 23, 2021

• final commit, demos are put in a queue, the idea is to feed these demos to the sampler and simply fill the RepayBuffer with them, for that however, the camera observations (and rewards? (and done?,…

  Show description for a0d570e
  deepdad

  committed
  a0d570e

  Copy full SHA for a0d570e

Commits on Jul 22, 2021

• refactor(README,-envs/rlbench.py,-main.py): adds task builder to README, corrects task name in main, changes camera in envs/rlbench to wrist

  Show description for ece5820
  deepdad

  committed
  ece5820

  Copy full SHA for ece5820

Commits on Jul 21, 2021

• refactor(envs/rlbench.py,-site-packages/../RLBENCH,-main.py): Enlarges the target object and deletes the distractors in RLBENCH reach_target.py task. Uses dense reward

  Show description for e207c73
  deepdad

  committed
  e207c73

  Copy full SHA for e207c73

Commits on Jul 20, 2021

• feat(test-for-the-reward-function:): Experiment of the reward function compared against sparse rewards. Some of these changes are also good for main branch (i.e. cleaning up the rlbench.py environm…

  Show description for e58fcb3
  deepdad

  committed
  e58fcb3

  Copy full SHA for e58fcb3

Commits on Jul 19, 2021

• feat(main.py): When running main.py with --eval, n_val_envs must be > 0. This minor change sets it to 1 automatically. However, this doesn't work, at least not when running on remote Desktop. Qt co…

  Show description for 1e792c7
  deepdad

  committed
  1e792c7

  Copy full SHA for 1e792c7

Commits on Jul 17, 2021

• build(requirements.txt): Adds requirements.txt Moves all the platform tests and basic DLLab files to folder test/

  Show description for a1fa53f
  deepdad

  committed
  a1fa53f

  Copy full SHA for a1fa53f

Commits on Jul 15, 2021

• code comments wrt filling the replay buffer

  Show description for 195a6d7
  deepdad

  committed
  195a6d7

  Copy full SHA for 195a6d7

Commits on Jul 12, 2021

• set batch size and length in main etc. (param. changes)

  deepdad

  committed
  ccb00e7

  Copy full SHA for ccb00e7

Commits on Jul 9, 2021

• disable 5/6 cameras

  deepdad

  committed
  d5de346

  Copy full SHA for d5de346

Commits on Jul 8, 2021

• maybe cast float to int for the box, but part of the atari2.py environment is really not useful and can be removed

  deepdad

  committed
  046683d

  Copy full SHA for 046683d

• partial gymRLBencRLBEnch gym based on RLPyt atari_env

  deepdad

  committed
  b698e35

  Copy full SHA for b698e35

• rlbench env

  deepdad

  committed
  aa5e38d

  Copy full SHA for aa5e38d

• pulling in the main from jh-branch

  deepdad

  committed
  111cd58

  Copy full SHA for 111cd58

• trying to get RLBench task to run on tfpool but fails on Qt application (CoppeliaSim) problem and GLib error

  deepdad

  committed
  0a70a4d

  Copy full SHA for 0a70a4d

Commits on Jul 7, 2021

• str object is not callable

  deepdad

  committed
  31794cc

  Copy full SHA for 31794cc

• changes dmc_dreamer_agent to benchmark_dreamer_agent

  deepdad

  committed
  a565442

  Copy full SHA for a565442

Commits on Jul 6, 2021

• changes to the env class due to errors with camera now its runnig with rl_bench_dreamer

  Wetzr

  committed
  a565222

  Copy full SHA for a565222

Commits on Jul 5, 2021

• added wrapperclass for rlbench (got error)

  Wetzr

  committed
  6f4166d

  Copy full SHA for 6f4166d

• test

  Wetzr

  committed
  3cce1dc

  Copy full SHA for 3cce1dc

Commits on Jul 1, 2021

• checks and tests

  deepdad

  committed
  c6747a2

  Copy full SHA for c6747a2

Loading