Bug in "Aarons_portfolio_optimization_example"

[gabrielp18]

When trying to run this part:

train_df = data_split(df, start=config.TRAIN_START_DATE,
                        end=config.TRAIN_END_DATE)
stock_dimension = len(train_df.tic.unique())
state_space = stock_dimension
print(f"Stock Dimension: {stock_dimension}, State Space: {state_space}")

env_kwargs = {
    "hmax": hmax,
    "initial_amount": initial_amount,
    "transaction_cost_pct": transaction_cost_pct,
    "state_space": state_space,
    "stock_dim": stock_dimension,
    "tech_indicator_list": config.INDICATORS,
    "action_space": stock_dimension,
    "reward_scaling": reward_scaling
}
# use cached model if you are iterating on evaluation code
if use_cached_model and os.path.exists("saved_models/a2c_model.pt"):
    trained_a2c = A2C.load("saved_models/a2c_model.pt")
else:
    e_train_gym = StockPortfolioEnv(df=train_df, **env_kwargs)
    agent = DRLAgent(env=e_train_gym)

    A2C_PARAMS = {
        "n_steps": 10,
        "ent_coef": 0.005,
        "learning_rate": 0.0004
    }
    model_a2c = agent.get_model(model_name="a2c", model_kwargs=A2C_PARAMS)
    trained_a2c = agent.train_model(
        model=model_a2c,
        tb_log_name='a2c',
        total_timesteps=40000
        )
    # save trained_a2c model
    trained_a2c.save("saved_models/a2c_model.pt")

# evaluate on test data
test_df = data_split(
    df,
    start=config.TEST_START_DATE,
    end=config.TEST_END_DATE
)
e_test_gym = StockPortfolioEnv(df=test_df, **env_kwargs)

df_daily_return, df_actions = DRLAgent.DRL_prediction(
    model=trained_a2c,
    environment=e_test_gym,
)
df_daily_return["method"] = "a2c"

I got the following error:

ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (2,) + inhomogeneous part.

[rajathk2003]

Are you still facing the issue or has this been fixed?