https://github.com/nixtla logo
k

Kaustav Chaudhury

04/03/2023, 12:38 PM
Hi what is tye difference between batch size and windows per batch size in TFT
c

Cristian (Nixtla)

04/03/2023, 1:57 PM
Hi @Kaustav Chaudhury! In every training iteration the data loader samples
batch_size
time-series from the dataset. Then, the model samples
windows_batch_size
individual windows (of size
input_size
+
h
) from the time-series in the batch.
Usually
windows_batch_size
is more important to tune since it ultimately determines the size of the batch for the model for gradient computation.
k

Kaustav Chaudhury

04/03/2023, 4:37 PM
So if I have window batch size of 128 per say it means it has 128 windows in each batch right
c

Cristian (Nixtla)

04/03/2023, 6:55 PM
Yes
3 Views