mirror of
https://github.com/vale981/ray
synced 2025-03-06 10:31:39 -05:00
43 lines
1.2 KiB
YAML
43 lines
1.2 KiB
YAML
|
# Note: HalfCheetahBulletEnv-v0 is not the same as MuJoCo's HalfCheetah-v0.
|
||
|
|
ddpg-halfcheetahbulletenv-v0:
|
||
|
|
env: HalfCheetahBulletEnv-v0
|
||
|
|
run: DDPG
|
||
|
|
stop:
|
||
|
|
episode_reward_mean: -300.0
|
||
|
|
timesteps_total: 200000
|
||
|
|
config:
|
||
|
|
actor_hiddens: [256, 256]
|
||
|
|
critic_hiddens: [256, 256]
|
||
|
|
n_step: 3
|
||
|
|
model: {}
|
||
|
|
gamma: 0.99
|
||
|
|
env_config: {}
|
||
|
|
exploration_config:
|
||
|
|
initial_scale: 1.0
|
||
|
|
final_scale: 0.02
|
||
|
|
scale_timesteps: 10000
|
||
|
|
ou_base_scale: 0.1
|
||
|
|
ou_theta: 0.15
|
||
|
|
ou_sigma: 0.2
|
||
|
|
timesteps_per_iteration: 1000
|
||
|
|
target_network_update_freq: 0
|
||
|
|
tau: 0.001
|
||
|
|
buffer_size: 15000
|
||
|
|
prioritized_replay: true
|
||
|
|
prioritized_replay_alpha: 0.6
|
||
|
|
prioritized_replay_beta: 0.4
|
||
|
|
prioritized_replay_eps: 0.000001
|
||
|
|
clip_rewards: false
|
||
|
|
actor_lr: 0.001
|
||
|
|
critic_lr: 0.001
|
||
|
|
use_huber: true
|
||
|
|
huber_threshold: 1.0
|
||
|
|
l2_reg: 0.000001
|
||
|
|
learning_starts: 500
|
||
|
|
rollout_fragment_length: 1
|
||
|
|
train_batch_size: 48
|
||
|
|
num_workers: 0
|
||
|
|
num_gpus: 1
|
||
|
|
num_gpus_per_worker: 0
|
||
|
|
worker_side_prioritization: false
|