Skip to content

Hyper parameters

Chris Rhodes edited this page Jun 15, 2019 · 3 revisions
Hyper-parameter name Default Value
batch_size 64
num_epochs 10
stack_size 1
lr 0.0003
exploration_type categorical
e_greedy_value 0.05
epsilon_steps 10000
beta_entropy 0.01
discount_factor 0.999
loss_type Mean squared error
num_episodes_between_training 20
term_cond_max_episodes 100000
term_cond_avg_score 100000

You can set the values in rl_deepracer_coach_robomaker.py, in the hyperparameters variable.

Clone this wiki locally