I'm currently training a RNNModel using python darts. For comparing different trained models I want to extract the train_loss and val_loss from the fit method. How would I do this? I've read something about a metric collection but can't figure out how to use it.
This is my current code
from darts.models import RNNModel
from darts import TimeSeries
train = # training data as TimeSeries
model = RNNModel(model="LSTM", input_chunk_length=self.past_samples)
model.fit(train)
The loss is shown in the console during training but I don't know how to access it.
So far I've tried to find any docs online and asked bing chat and ChatGPT for help. However they told me to use model.history.history["loss"] which doesn't exist