| from easydict import EasyDict | |
| import ding.envs.gym_env | |
| cfg = dict( | |
| exp_name='lunarlander_c51', | |
| seed=0, | |
| env=dict( | |
| collector_env_num=8, | |
| evaluator_env_num=8, | |
| env_id='LunarLander-v2', | |
| n_evaluator_episode=8, | |
| stop_value=260, | |
| ), | |
| policy=dict( | |
| cuda=False, | |
| model=dict( | |
| obs_shape=8, | |
| action_shape=4, | |
| encoder_hidden_size_list=[512, 64], | |
| v_min=-30, | |
| v_max=30, | |
| n_atom=51, | |
| ), | |
| discount_factor=0.99, | |
| nstep=3, | |
| learn=dict( | |
| update_per_collect=10, | |
| batch_size=64, | |
| learning_rate=0.001, | |
| target_update_freq=100, | |
| ), | |
| collect=dict( | |
| n_sample=64, | |
| unroll_len=1, | |
| ), | |
| other=dict( | |
| eps=dict( | |
| type='exp', | |
| start=0.95, | |
| end=0.1, | |
| decay=50000, | |
| ), replay_buffer=dict(replay_buffer_size=100000, ) | |
| ), | |
| ), | |
| wandb_logger=dict( | |
| gradient_logger=True, video_logger=True, plot_logger=True, action_logger=True, return_logger=False | |
| ), | |
| ) | |
| cfg = EasyDict(cfg) | |
| env = ding.envs.gym_env.env | |