Yearly seasonality with limited data set

Hello Everyone,
I am presently working on electricity demand forecasting model(s) using facebook prophet. For certain buildings we only have data from 2020 onwards with yearly seasonality. However the challenge is that with covid the demand on those buildings for 2020 were lower compared to the 2021 and hence the model is not exactly fitting the actual.
m= Prophet(holidays=holidays,yearly_seasonality=True,seasonality_mode="multiplicative",changepoint_prior_scale=0.01,seasonality_prior_scale=0.01)
m.add_regressor('mean external temperature')
m.add_regressor('weekday_N')
m.add_regressor('weekday_Y')
m.fit(prophet_training_set)
test_forecast = m.predict(prophet_test_set)
Is there any guidance on how to deal with situation