Problem with global model

[MamyGhost]

I have testing to train with global and local modele and I have a huge diference in the result , local model fit better than global model
what could be the reason for this big difference?

global :
trend_global_local="local", season_global_local="local", yearly_seasonality=True, weekly_seasonality=False, daily_seasonality=False, impute_missing=True, n_lags=30*3, n_forecasts=365, growth="discontinuous", epochs=200

local:
yearly_seasonality=True, weekly_seasonality=False, daily_seasonality=False, impute_missing=True, n_lags=30*3, n_forecasts=365, growth="discontinuous", epochs=200
@ourownstory