reward_signal_steps_per_update: 10.0
encoder_layers: 1
policy_layers: 1
forward_layers: 1
forward_layers: 0
value_layers: 1
feature_size: 16
# separate_value_net: true
train_model: false
load_policy: true
load_value: true
transfer_path: "results/sac-ball-f16-e1p1f1/3DBall"
transfer_path: "results/sac-ball-f16-e1p1f0/3DBall"
network_settings:
normalize: true
hidden_units: 128