Date and Time:
17/09/2021
15:37:41

env: Walker2dBackward-v1
hrl: False
rl_num_hid_layers: 2
rl_hid_size: 32
rl_activation: tanh
rl_method: trpo
rl_fixed_var: True
meta_duration: 30
meta_oracle: True
meta_num_hid_layers: 2
meta_hid_size: 32
meta_activation: tanh
meta_max_grad_norm: 10.0
meta_method: ppo
meta_entcoeff: 0.0002
use_trans: True
use_trans_between_same_policy: False
trans_term_activation: softmax
trans_term_prob: 0.02
trans_duration: 100
trans_num_hid_layers: 2
trans_hid_size: 32
trans_activation: tanh
trans_max_grad_norm: 10.0
trans_method: ppo
trans_fixed_var: True
trans_entcoeff: 0.001
trans_include_acc: True
trans_include_task_obs: False
trans_apply_first_time_step: False
use_proximity_predictor: False
proximity_loss_type: lsgan
proximity_use_traj_portion_start: ['0']
proximity_use_traj_portion_end: ['1']
proximity_num_hid_layers: 2
proximity_hid_size: 96
proximity_activation_fn: relu
proximity_learning_rate: 0.0001
proximity_optim_epochs: 5
proximity_hist: False
proximity_hist_num_bin: 10
proximity_dense_diff_rew: True
proximity_dense_diff_rew_final_bonus: True
proximity_include_acc: True
proximity_obs_norm: True
proximity_only_use_trans_term_state: False
proximity_replay_size: 1000000
proximity_keep_collected_obs: True
proximity_weight_decay_linear: False
proximity_weight_decay_rate: 0.95
proximity_max_grad_norm: 10.0
primitive_envs: None
primitive_dir: ./log
primitive_paths: None
primitive_num_hid_layers: 2
primitive_hid_size: 32
primitive_activation: tanh
primitive_method: trpo
primitive_fixed_var: True
primitive_include_acc: False
primitive_use_term: True
is_train: True
load_meta_path: None
load_model_path: None
write_summary_step: 5
ckpt_save_step: 100
max_iters: 10001
num_rollouts: 10000
num_batches: 32
num_trans_batches: 256
num_evaluation_run: 10
evaluate_proximity_predictor: False
evaluate_all_ckpts: False
evaluation_log: True
video_caption_off: False
final_eval: False
test_module_net: False
is_collect_state: False
lr_decay: True
clip_param: 0.2
entcoeff: 0.0
optim_epochs: 10
optim_stepsize: 0.0001
optim_batchsize: 64
max_kl: 0.01
cg_iters: 10
cg_damping: 0.1
vf_stepsize: 0.001
vf_iters: 5
prefix: backward_ICLR2019_seed123
render: False
record: False
video_prefix: None
log_dir: log/Walker2dBackward.backward_ICLR2019_seed123
seed: 123
debug: False
env_args_str: None

Environment argument:
alive_reward: 1
angle_reward: 0.1
apply_force: 100
ctrl_reward: 0.001
foot_reward: 0.01
height_reward: 1
init_randomness: 0.01
min_height: 0.8
prob_apply_force: 0.2
random_steps: 5
x_vel_limit: 3
x_vel_reward: 1
