Training NBEATSEnsembleEstimator with multiple time series #1347
Unanswered
praveen049
asked this question in
Q&A
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hi
I am trying to train with NBEATSEnsembleEstimator a training dataset containing over 100 time series.
To get the implementation right before beginning parameter tuning - i am training with just 5 epochs and 10 batches per epoch.
Below is the estimator.
estimator = NBEATSEnsembleEstimator( prediction_length=prediction_length, freq="H", trainer=Trainer(ctx="cpu", learning_rate=0.001, epochs=5, num_batches_per_epoch=20, batch_size=16) )
I want to know if training is 5 epochs for each time series ? That is, are time series fit sequentially one at a time ?
Thanks
Beta Was this translation helpful? Give feedback.
All reactions