File size: 2,079 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 66 67 68 69 70 71 72 |
from easydict import EasyDict
halfcheetah_bdq_config = dict(
exp_name='halfcheetah_bdq_seed0',
env=dict(
env_id='HalfCheetah-v3',
norm_reward=dict(use_norm=False, ),
collector_env_num=8,
evaluator_env_num=8,
n_evaluator_episode=8,
stop_value=12000,
action_bins_per_branch=2,
),
policy=dict(
cuda=False,
priority=False,
discount_factor=0.99,
nstep=1,
model=dict(
obs_shape=17,
num_branches=6,
action_bins_per_branch=2, # mean the action shape is 6, 2 discrete actions for each action dimension
encoder_hidden_size_list=[256, 256, 128],
),
learn=dict(
batch_size=512,
learning_rate=3e-4,
ignore_done=True,
target_update_freq=500,
update_per_collect=20,
),
collect=dict(
n_sample=256,
unroll_len=1,
),
eval=dict(evaluator=dict(eval_freq=1000, )),
other=dict(
# Epsilon greedy with decay.
eps=dict(
# Decay type. Support ['exp', 'linear'].
type='exp',
start=1,
end=0.05,
decay=int(1e5),
),
replay_buffer=dict(replay_buffer_size=int(1e6), )
),
),
)
halfcheetah_bdq_config = EasyDict(halfcheetah_bdq_config)
main_config = halfcheetah_bdq_config
halfcheetah_bdq_create_config = dict(
env=dict(
type='mujoco',
import_names=['dizoo.mujoco.envs.mujoco_env'],
),
env_manager=dict(type='subprocess'),
policy=dict(type='bdq', ),
)
halfcheetah_bdq_create_config = EasyDict(halfcheetah_bdq_create_config)
create_config = halfcheetah_bdq_create_config
if __name__ == "__main__":
# or you can enter `ding -m serial_onpolicy -c halfcheetah_onbdq_config.py -s 0`
from ding.entry import serial_pipeline
serial_pipeline(
(main_config, create_config),
seed=0,
max_env_step=10000000,
)
|