rail-berkeley/rlkit issues and pull requests

#48 - Make sure you dont need a Mujoco license to use any of the algorithms

Issue - State: closed - Opened by redknightlois over 5 years ago - 5 comments

#47 - Numpy to Pytorch should ignore Pytorch Tensors

Pull Request - State: closed - Opened by redknightlois over 5 years ago - 5 comments

#46 - HOME does not exist on Windows

Pull Request - State: closed - Opened by redknightlois over 5 years ago - 2 comments

#45 - while pass ‘render=True’ get 'Window rendering not supported'

Issue - State: closed - Opened by cww97 over 5 years ago - 2 comments

#44 - how can I render while training

Issue - State: closed - Opened by cww97 over 5 years ago - 1 comment

#43 - Future entropy missed in SAC.

Issue - State: closed - Opened by brickerino over 5 years ago - 4 comments
Labels: bug

#42 - Add online rl algorithm + train_mode function

Pull Request - State: closed - Opened by hexiang-hu over 5 years ago - 3 comments

#41 - Performance on Hopper-v2

Issue - State: closed - Opened by quanvuong over 5 years ago - 11 comments

#40 - Value network in TwinSAC

Issue - State: closed - Opened by jendelel over 5 years ago - 3 comments

#39 - snapshot model weights and optionally load pre-trained weights before training

Pull Request - State: closed - Opened by katerakelly over 5 years ago

#38 - The policy loss in SAC?

Issue - State: closed - Opened by zhaoyingnan179346 over 5 years ago - 1 comment

#37 - Use stable implementation of tanh's inverse log determinant Jacobian

Issue - State: closed - Opened by alexlee-gk over 5 years ago - 1 comment

#36 - Training env is reset at start of epoch, but PathBuilder not reset. Resulting in erroneous Exploration statistics.

Issue - State: closed - Opened by pimdh almost 6 years ago - 4 comments
Labels: bug

#35 - Features/openai hacks

Pull Request - State: open - Opened by richardrl almost 6 years ago - 1 comment

#34 - ImportError: No module named 'glfw'

Issue - State: open - Opened by tseyde almost 6 years ago - 8 comments

#33 - best way to resume training from PKL

Issue - State: open - Opened by richardrl almost 6 years ago - 11 comments

#32 - HER DQN enabled, along with example code running on gridworld.

Pull Request - State: closed - Opened by cdevin almost 6 years ago

#31 - Cannot learn (pusher experiment)

Issue - State: open - Opened by Nicolas99-9 almost 6 years ago - 10 comments

#30 - Unable to reproduce the results of HER-TD3

Issue - State: closed - Opened by charliezon almost 6 years ago - 6 comments
Labels: bug

#29 - removed extraneous full_observation appendage

Pull Request - State: closed - Opened by richardrl almost 6 years ago

#28 - Support discrete action in HER replay buffer

Issue - State: closed - Opened by vitchyr almost 6 years ago - 1 comment
Labels: bug, enhancement

#27 - Multiple worker support

Issue - State: open - Opened by richardrl almost 6 years ago - 10 comments
Labels: enhancement

#26 - Retrieve s3

Issue - State: closed - Opened by richardrl almost 6 years ago - 1 comment

#25 - Making local mode Not run through doodad

Issue - State: closed - Opened by richardrl almost 6 years ago - 2 comments

#24 - Mujoco_py doesn't build

Issue - State: closed - Opened by richardrl almost 6 years ago - 6 comments

#23 - HER example scripts segfault when running "local_docker" or "ec2" mode

Issue - State: closed - Opened by richardrl almost 6 years ago - 3 comments

#22 - Hotfixes/unicodeerror

Pull Request - State: open - Opened by richardrl almost 6 years ago - 2 comments

#21 - HER/RIG example scripts expect "scripts" folder to be in "rlkit/scripts"

Issue - State: closed - Opened by richardrl almost 6 years ago - 2 comments

#20 - catch unbound localerror

Pull Request - State: open - Opened by richardrl almost 6 years ago - 2 comments

#19 - How could I see the video in rig...

Issue - State: closed - Opened by cww97 almost 6 years ago - 2 comments

#18 - viskit incompatability: plotly.exceptions.PlotlyDictValueError: 'title' has invalid value inside 'layout'

Issue - State: closed - Opened by vitchyr almost 6 years ago - 1 comment
Labels: bug

#17 - Correct full name of TD3

Pull Request - State: closed - Opened by seungjaeryanlee almost 6 years ago - 1 comment

#16 - ImportError: No module named 'rlkit'

Issue - State: closed - Opened by cww97 almost 6 years ago - 2 comments

#15 - Remove incorrect update date from readme

Issue - State: closed - Opened by impredicative almost 6 years ago - 1 comment

#14 - Update SAC url in README

Pull Request - State: closed - Opened by hartikainen almost 6 years ago

#13 - examples/tsac.py does not run at master

Issue - State: closed - Opened by deq2 about 6 years ago - 1 comment

#12 - AssertionError: Table keys cannot change from iteration to iteration

Issue - State: closed - Opened by vitchyr about 6 years ago - 1 comment
Labels: bug

#11 - Box2D-kengz installation crashing

Issue - State: closed - Opened by vitchyr about 6 years ago - 2 comments
Labels: bug

#10 - License

Issue - State: closed - Opened by yang-song over 6 years ago - 1 comment

#9 - Docker Image Does Not Work

Issue - State: closed - Opened by avaziri over 6 years ago - 2 comments

#8 - Getting a gtimer error while installing requirements

Issue - State: closed - Opened by sir-avinash over 6 years ago - 3 comments

#7 - How do we set reward scale for SAC

Issue - State: closed - Opened by qizhg over 6 years ago - 1 comment

#6 - Unable to Recreate TDM Performance for Half Cheetah

Issue - State: closed - Opened by Random-Word over 6 years ago - 3 comments

#5 - AttributeError: attribute 'site_pos' of 'mujoco_py.cymj.PyMjModel' objects is not writable

Issue - State: closed - Opened by jiameij over 6 years ago - 3 comments

#4 - TDM how to load trained weight for continue train

Issue - State: closed - Opened by zdx3578 over 6 years ago - 2 comments

GitHub / rail-berkeley/rlkit issues and pull requests