|
from easydict import EasyDict |
|
|
|
cartpole_impala_config = dict( |
|
exp_name='cartpole_impala_seed0', |
|
env=dict( |
|
collector_env_num=8, |
|
evaluator_env_num=5, |
|
n_evaluator_episode=5, |
|
stop_value=195, |
|
), |
|
policy=dict( |
|
cuda=False, |
|
model=dict( |
|
obs_shape=4, |
|
action_shape=2, |
|
encoder_hidden_size_list=[64, 64], |
|
), |
|
|
|
unroll_len=8, |
|
learn=dict( |
|
|
|
|
|
update_per_collect=4, |
|
|
|
batch_size=16, |
|
learning_rate=0.0005, |
|
|
|
value_weight=0.5, |
|
|
|
entropy_weight=0.0001, |
|
|
|
discount_factor=0.9, |
|
|
|
lambda_=0.95, |
|
|
|
rho_clip_ratio=1.0, |
|
|
|
c_clip_ratio=1.0, |
|
|
|
rho_pg_clip_ratio=1.0, |
|
), |
|
collect=dict( |
|
|
|
n_sample=16, |
|
|
|
discount_factor=0.9, |
|
gae_lambda=0.95, |
|
collector=dict(collect_print_freq=1000, ), |
|
), |
|
eval=dict(evaluator=dict(eval_freq=200, )), |
|
other=dict(replay_buffer=dict( |
|
replay_buffer_size=1000, |
|
max_use=16, |
|
), ), |
|
), |
|
) |
|
|
|
cartpole_impala_config = EasyDict(cartpole_impala_config) |
|
main_config = cartpole_impala_config |
|
|
|
cartpole_impala_create_config = dict( |
|
env=dict( |
|
type='cartpole', |
|
import_names=['dizoo.classic_control.cartpole.envs.cartpole_env'], |
|
), |
|
env_manager=dict(type='base'), |
|
policy=dict(type='impala'), |
|
) |
|
|
|
cartpole_impala_create_config = EasyDict(cartpole_impala_create_config) |
|
create_config = cartpole_impala_create_config |
|
|
|
if __name__ == "__main__": |
|
|
|
from ding.entry import serial_pipeline |
|
serial_pipeline((main_config, create_config), seed=0) |
|
|