use_budget_control=use_budget_control,
use_prioritized_experience_replay=use_prioritized_replay,
max_trajectory_length=user_max_request_time,
update_times_per_train=update_times_per_train)
if train:
run_env(agent=agent,
user_num=user_num,
training_episode=1000,
training_log_interval=1,
test_interval_list=[100, 5],
test_round=1,
seed=seed,
init_roi_th=init_cpr_thr,
use_prioritized_replay=use_prioritized_replay,
budget=budget,
use_budget_control=use_budget_control,
user_max_request_time=user_max_request_time)
else:
eval_env(
agent,
"./result/20200206_learn_multiseed_result/MSBCB/train/1000_user/action_n=2/per=0/seed=1/1580959853/best_model/model-1000",
user_num=user_num,
seed=seed,
test_epoch=1,
init_roi_th=init_cpr_thr,
print_log=False,
use_prioritized_replay=use_prioritized_replay,
budget=budget,
use_budget_control=use_budget_control,
user_max_request_time=user_max_request_time)
update_times_per_train=update_times_per_train)
if train:
run_env(
agent=agent,
user_num=user_num,
training_episode=200,
training_log_interval=1,
test_interval_list=[100, 5],
# user_num=user_num, training_episode=1000, training_log_interval=1, test_interval_list=[100, 5],
test_round=1,
seed=seed,
init_roi_th=init_cpr_thr,
use_prioritized_replay=use_prioritized_replay,
budget=budget,
use_budget_control=use_budget_control,
user_max_request_time=user_max_request_time)
else:
eval_env(
agent,
"./exp/learning_result/GreedyDQN/train/1000_user/action_n=11/per=0/seed=1/1581679892/best_model/model-235",
user_num=user_num,
seed=seed,
test_epoch=10,
init_roi_th=init_cpr_thr,
print_log=False,
use_prioritized_replay=use_prioritized_replay,
budget=budget,
use_budget_control=use_budget_control,
user_max_request_time=user_max_request_time)
use_budget_control=use_budget_control,
use_prioritized_experience_replay=use_prioritized_replay,
max_trajectory_length=7,
update_times_per_train=update_times_per_train)
if train:
run_env(agent=agent,
user_num=user_num,
training_episode=500 * 2,
training_log_interval=1,
test_interval_list=[100, 5],
test_round=1,
seed=seed,
init_roi_th=init_cpr_thr,
use_prioritized_replay=use_prioritized_replay,
budget=budget,
use_budget_control=use_budget_control,
user_max_request_time=7)
else:
eval_env(
agent,
"result/1000_user_no_roi_thr/GreedyDDPG/train/1000_user/action_n=1/per=0/seed=1/1578469558/best_model/model-800",
user_num=user_num,
seed=seed,
test_epoch=1,
init_roi_th=init_cpr_thr,
print_log=False,
use_prioritized_replay=use_prioritized_replay,
budget=budget,
use_budget_control=use_budget_control,
user_max_request_time=7)