Error with TRAIN and TEST interval about finrl-meta HOT 4 CLOSED

Thanks for reporting it. We are looking into it.

from finrl-meta.

setting TRAIN_END_DATE = '2021-09-30' also gives the below error

AttributeError                            Traceback (most recent call last)
<ipython-input-97-21c42d37b7d1> in <module>()
     11       erl_params=ERL_PARAMS,
     12       break_step=5e4,
---> 13       if_vix=False
     14       )

3 frames
<ipython-input-36-a440fe771f96> in train(start_date, end_date, ticker_list, data_source, time_interval, technical_indicator_list, drl_lib, env, model_name, if_vix, **kwargs)
     35         trained_model = agent.train_model(model=model, 
     36                                           cwd=current_working_dir,
---> 37                                           total_timesteps=break_step)
     38 
     39     elif drl_lib == 'rllib':

/FinRL-Meta/drl_agents/elegantrl_models.py in train_model(self, model, cwd, total_timesteps)
     79         model.cwd = cwd
     80         model.break_step = total_timesteps
---> 81         train_and_evaluate(args=model)
     82 
     83     @staticmethod

/usr/local/lib/python3.7/dist-packages/elegantrl/train/run_tutorial.py in train_and_evaluate(***failed resolving arguments***)
     91             traj_list = agent.explore_env(env, target_step)
     92             steps, r_exp = update_buffer(traj_list)
---> 93         logging_tuple = agent.update_net(buffer, batch_size, repeat_times, soft_update_tau)
     94         with torch.no_grad():
     95             if_reach_goal, if_save = evaluator.evaluate_and_save(agent.act, steps, r_exp, logging_tuple)

/usr/local/lib/python3.7/dist-packages/elegantrl/agents/AgentPPO.py in update_net(self, buffer, batch_size, repeat_times, soft_update_tau)
    196 
    197         a_std_log = getattr(self.act, 'a_std_log', torch.zeros(1)).mean()
--> 198         return obj_critic.item(), obj_actor.item(), a_std_log.item()  # logging_tuple
    199 
    200     def get_reward_sum_raw(self, buf_len, buf_reward, buf_mask, buf_value) -> (torch.Tensor, torch.Tensor):

AttributeError: 'NoneType' object has no attribute 'item'

from finrl-meta.

setting TRAIN_END_DATE = '2021-09-30' also gives the below error

AttributeError                            Traceback (most recent call last)
<ipython-input-97-21c42d37b7d1> in <module>()
     11       erl_params=ERL_PARAMS,
     12       break_step=5e4,
---> 13       if_vix=False
     14       )

3 frames
<ipython-input-36-a440fe771f96> in train(start_date, end_date, ticker_list, data_source, time_interval, technical_indicator_list, drl_lib, env, model_name, if_vix, **kwargs)
     35         trained_model = agent.train_model(model=model, 
     36                                           cwd=current_working_dir,
---> 37                                           total_timesteps=break_step)
     38 
     39     elif drl_lib == 'rllib':

/FinRL-Meta/drl_agents/elegantrl_models.py in train_model(self, model, cwd, total_timesteps)
     79         model.cwd = cwd
     80         model.break_step = total_timesteps
---> 81         train_and_evaluate(args=model)
     82 
     83     @staticmethod

/usr/local/lib/python3.7/dist-packages/elegantrl/train/run_tutorial.py in train_and_evaluate(***failed resolving arguments***)
     91             traj_list = agent.explore_env(env, target_step)
     92             steps, r_exp = update_buffer(traj_list)
---> 93         logging_tuple = agent.update_net(buffer, batch_size, repeat_times, soft_update_tau)
     94         with torch.no_grad():
     95             if_reach_goal, if_save = evaluator.evaluate_and_save(agent.act, steps, r_exp, logging_tuple)

/usr/local/lib/python3.7/dist-packages/elegantrl/agents/AgentPPO.py in update_net(self, buffer, batch_size, repeat_times, soft_update_tau)
    196 
    197         a_std_log = getattr(self.act, 'a_std_log', torch.zeros(1)).mean()
--> 198         return obj_critic.item(), obj_actor.item(), a_std_log.item()  # logging_tuple
    199 
    200     def get_reward_sum_raw(self, buf_len, buf_reward, buf_mask, buf_value) -> (torch.Tensor, torch.Tensor):

AttributeError: 'NoneType' object has no attribute 'item'

You may try to reset 'target_step' as a number larger than the episode steps. You can set target_step by setting ERL_PARAMS = {..., "target_step": x ,...} .

from finrl-meta.

@rayrui312 Yep you are right.
Regarding the original issue, i came across it too but it works on retrying. probably just an issue with requests

from finrl-meta.