jkulhanek / robot-visual-navigation

Visual Navigation in Real-World Indoor Environments Using End-to-End Deep Reinforcement Learning Official Implementation
MIT License
58 stars 11 forks source link

puzzle about training result #16

Closed GongYanfu closed 1 year ago

GongYanfu commented 2 years ago

Is this training result normal? espacially for the result of "turtlebot" and "turtlebot-a2c" image

jkulhanek commented 2 years ago

Not sure, it doesn’t look like my training results.

GongYanfu commented 2 years ago

Can you show the training result of this? about experiments of turtlebot, turtlebot-a2c and turtlebot-unreal

jkulhanek commented 2 years ago

I believe it is in the paper.


From: GongYanfu @.> Sent: Tuesday, September 6, 2022 1:00:52 PM To: jkulhanek/robot-visual-navigation @.> Cc: Jonáš Kulhánek @.>; Comment @.> Subject: Re: [jkulhanek/robot-visual-navigation] puzzle about training result (Issue #16)

Can you show the training result of this? about experiments of turtlebot, turtlebot-a2c and turtlebot-unreal

— Reply to this email directly, view it on GitHubhttps://github.com/jkulhanek/robot-visual-navigation/issues/16#issuecomment-1238050542, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABAGZIXXTPZTLLL5EE6MLR3V44W7JANCNFSM54O2CFEQ. You are receiving this because you commented.Message ID: @.***>

jkulhanek commented 1 year ago

This should now be resolved by commit: https://github.com/jkulhanek/robot-visual-navigation/commit/2c87a2d871eb2982a1070fe4425bd1218d4958b8