Describe the task. It can be a feature, documentation, etc.

We need to implement a transformer based predictor for JEPA, following the literature (see https://www.overleaf.com/project/68dd158b1e327e168d9ceebd)

We will use the existing attention blocks and MLP layers for this. This resembles the forecasting engine architecture. A quick look revelas that it is not very straight forward to refactor these two to a single class, so first we will implement it as a seperate class.

Hedgedoc URL, if you are keeping notes, plots, logs in hedgedoc.

No response

Area

• datasets, data readers, data preparation and transfer
• model
• science
• infrastructure and engineering
• evaluation, export and visualization
• documentation