Skip to content

Latest commit

 

History

History
31 lines (17 loc) · 1.22 KB

README.md

File metadata and controls

31 lines (17 loc) · 1.22 KB

Baselines

RLBaselines is built upon OpenAI Baselines and uses improved documentation and code quality from Stable Baselines. Follow OpenAI's installation guide to get started.

Improved Plotting

  1. Added tensorboard-aggregator to aggregate multiple tensorboard runs.
  2. Added csv_aggregator.py for easier and more intuitive plotting of mean and standard deviation.

The shaded blue is the standard deviation from the mean, while the orange shade is the error in estimate of the mean.

image

  1. Added csv_plot.py to improve ease of use in utilising baseline plotting functions.

image

Results

image

Road map

  • Plot results using seaborn
  • Compare results in a more professional manner
  • Implement Prioritized Sequence Experience Replay
  • Implement Energy Based Prioritization
  • Compare the above