Denys88 / rl_games

RL implementations
MIT License
800 stars 135 forks source link

How to understand the swap_and_flattern01() operation #274

Closed BadWomanY closed 4 months ago

BadWomanY commented 4 months ago

I see that in a2c_common.py for the play_steps_rnn file, after collecting 16 eval observations and feeding them into the experience buffer, why do we need to apply swap_and_flattern01 to get_transformed_list function to modify the experience buffer? I am confused about this operation here.

BadWomanY commented 4 months ago

Actually I found the reason! All good now!