[Core] CV fails when df is not provided #800
Description
What happened + What you expected to happen
This line
neuralforecast/neuralforecast/core.py
Line 542 in ff1566c
tries to merge with
df, which can be None, so in that case it fails.
Originally reported here
Versions / Dependencies
1.6.4
Reproduction script
model = NeuralForecast(
[AutoNBEATS(**model_args)],
freq='D',
)
model.fit(df=data, val_size=horizon)
crossvalidation_df = model.cross_validation()
Issue Severity
Low: It annoys or frustrates me.