Skip to content

Latest commit

 

History

History
12 lines (9 loc) · 603 Bytes

README.md

File metadata and controls

12 lines (9 loc) · 603 Bytes
Raw

Behavior_Cloning

Behaviour Clonning On OpenAI Environment

DAgger Algorithm For Behavior Cloning In Open AI Lunar Lander

  1. Train Behavioral Policy Network from human data or any perfect System ( We used pre trained Nural Network Model of Lunar Lander as our Master Network) D_collection
  2. Re-run the Behavioural Policy network on new simulation and colect the data.
  3. Correctly label the generated data
  4. Aggrigate the generated and corrected data with the D_collection and retrain
  5. Repeat step 2-5 and measure the performance

Special Thanks to : @nikhilbarhate99 for his PPO Model