)
env = Learner(
log_dir=log_dir,
i_data=data[1]['train'],
j_data=data[0]['train'],
tst_data=data[1]['tst'],
learning_steps=500,
learning_steps_max=50000,
)
with open(file_name, 'a') as file:
file.write('\n\n===NEW DATA LOADING====\n\n')
file.write('\n\ni_data: ' + str(1))
file.write('\n\nj_data: ' + str(3))
for i in range(5):
train_DDQN(RL=RL_prio, env=env, file_name=file_name, penalty=0.02)
# double_DQN.plot_cost()
j = np.random.randint(4)
env.i_data = data[j]['train']
env.reward_tst_data = data[j]['tst']
k = np.random.randint(4)
env.j_data = data[k]['train']
with open(file_name, 'a') as file:
file.write('\n\n===NEW DATA LOADING====\n\n')
file.write('\n\ni_data: ' + str(j))
file.write('\n\nj_data: ' + str(k))
double_DQN.plot_cost()
double_DQN.save_model('episode-3')
env.destroy()
