Increasing training time of a2c #140
Description
I use the example of pong_a2c.py to train on my own env, and I find that the training time increasing after each epoch?
Is there any way to improve it?
{{ message }}
I use the example of pong_a2c.py to train on my own env, and I find that the training time increasing after each epoch?
Is there any way to improve it?