File size: 1,984 Bytes
079c32c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 |
from easydict import EasyDict
gym_hybrid_hppo_config = dict(
exp_name='gym_hybrid_hppo_seed0',
env=dict(
collector_env_num=8,
evaluator_env_num=5,
# (bool) Scale output action into legal range, usually [-1, 1].
act_scale=True,
env_id='Moving-v0', # ['Sliding-v0', 'Moving-v0']
n_evaluator_episode=5,
stop_value=1.8,
),
policy=dict(
cuda=True,
action_space='hybrid',
recompute_adv=True,
model=dict(
obs_shape=10,
action_shape=dict(
action_type_shape=3,
action_args_shape=2,
),
action_space='hybrid',
encoder_hidden_size_list=[256, 128, 64, 64],
sigma_type='fixed',
fixed_sigma_value=0.3,
bound_type='tanh',
),
learn=dict(
epoch_per_collect=10,
batch_size=320,
learning_rate=3e-4,
entropy_weight=0.5,
adv_norm=True,
value_norm=True,
),
collect=dict(
n_sample=3200,
discount_factor=0.99,
gae_lambda=0.95,
collector=dict(collect_print_freq=1000, ),
),
eval=dict(evaluator=dict(eval_freq=200, ), ),
),
)
gym_hybrid_hppo_config = EasyDict(gym_hybrid_hppo_config)
main_config = gym_hybrid_hppo_config
gym_hybrid_hppo_create_config = dict(
env=dict(
type='gym_hybrid',
import_names=['dizoo.gym_hybrid.envs.gym_hybrid_env'],
),
env_manager=dict(type='subprocess'),
policy=dict(type='ppo'),
)
gym_hybrid_hppo_create_config = EasyDict(gym_hybrid_hppo_create_config)
create_config = gym_hybrid_hppo_create_config
if __name__ == "__main__":
# or you can enter `ding -m serial -c gym_hybrid_hppo_config.py -s 0`
from ding.entry import serial_pipeline_onpolicy
serial_pipeline_onpolicy([main_config, create_config], seed=0, max_env_step=int(1e7))
|