plot the Training-Losses with the time series library "Darts"

Question

Is there a way to plot the Training-Losses from a model, which is trained with the time series library "Darts"? i am working with RNNs from the Darts-library for TimeSeries Forecasting. i am wondering, if there is a way to plot the training losses over the epochs.

with keras i do something like this:

model = Sequential() 
model.add(LSTM(100, activation='relu', input_shape=(n_input, n_features))) 
model.add(Dense(1)) 
model.compile(optimizer='adam', loss='mse') 

loss_per_epoch = model.history.history['loss'] plt.plot(range(len(loss_per_epoch)),loss_per_epoch) 
plt.show() 

maybe there is a way via tensorboard?

Answer 1

It looks like you can wrap your data in a darts.TimeSeries object, and then access the plotting from there. E.g.:

history_series = darts.TimeSeries.from_values(loss_per_epoch)
history_series.plot()