3DBall:
trainer_type: ppo_transfer
hyperparameters:
batch_size: 64
batch_size: 1200
buffer_size: 12000
learning_rate: 0.0003
beta: 0.001
gamma: 0.99
strength: 1.0
keep_checkpoints: 5
max_steps: 500000
max_steps: 1000000
time_horizon: 1000
summary_freq: 12000
threaded: true