I’m using GluonTS to make time series predictions with the TransformerPredictor. However, when I try to convert the forecast_it iterator to a list, I encounter the following error:

Setting an array element with a sequence. The requested array has an inhomogeneous shape after 2 dimensions. The detected shape was (15, 11) + inhomogeneous part. During handling of the above exception, another exception occurred: source array must be array like object

train_data = ListDataset(
    [{
        'start': df.index[0],
        'target': df['target'].values,
        'feat_dynamic_real': df['feat_dynamic_real'].values.T,
        'feat_static_cat': df['feat_static_cat'].values if 'feat_static_cat' in df.columns else None
    }],
    freq='D'
)

predictor = TransformerPredictor(
    input_size=1,
    prediction_length=30,
    freq='D',
    context_length=30,
    trainer=Trainer(epochs=10)
)

predictor.train(train_data)

forecast_it, ts_it = make_evaluation_predictions(dataset=train_data, predictor=predictor)

forecasts = list(forecast_it)  # This line raises the error

Tried ensuring all the time series are of equal lengths. Did not resolve the issue.