ValueError: cannot copy sequence with size 176 to array axis with dimension 181

[zcy-4747]

I wonder how 181 was calculated?

[zcy-4747]

I tried to use this model with other stock data, but failed

[emjay032]

maybe a little late, but 181 is the dimension of the state space:

balance + num_stocks+holdings +technichal indiactors
1+30+30+4*30 (4 technical indicators for each 30 stocks) = 181
balance = [balance]
num_stcosk = [stock_1,stock_2+...+stock_30]
technicahl indicators = [TI_1,stock_1,TI_2,stock_1,TI_3,stock_1,TI_4_stock_1,+...+TI_1,stock_30,TI_2,stock_30,TI_3,stock_30,TI_4_stock_30]
if you have a state space of dimension 176 change self.observation_space = spaces.Box(low=0, high=np.inf, shape = (181,)) to self.observation_space = spaces.Box(low=0, high=np.inf, shape = (176,))

low depends on the input of your data if you have negartive values i would recomend low = -np.inf

additionally when training and testing you should normalize the input to[-1,1] for your neural net by VecNormalize follwoing the Stable Basleine Documentation (training also the reward)
https://stable-baselines.readthedocs.io/en/master/guide/vec_envs.html