LSTM Time Series Regression - How does Pytorch train on multiple batches?

[AFordon]

Hello,

I’m following along with the Pytorch Time Series Regression (TSR) example and this article:

• Pytorch TSR Example
• Toward Data Science TSR Example

I would like more insight into how Pytorch trains on multiple sequences. It seems that the batches are trained in parallel, so how does loss and backpropagation get calculated? Will the individual batches receive different models (weights/bias)?