implement better ways of comparing algorithms

lucasosouza / fastRL

Library for Reinforcement Learning Algorithms, based on openAI environment implementation. Under development

MIT License

1 stars 0 forks source link

Open lucasosouza opened 6 years ago

lucasosouza commented 6 years ago

notebooks are not efficient to compare:

look into saving images, to visualize plots side by side, useful for reports and papers
numerical comparison and outputs (example confusion matrix in SL).

need better ways for evaluating different methods in different environments