Description of changes: This splits input_transform into context_input_transform and label_input_transform. Previously, input_transform was being used for both context and label during training which would lead to incorrect results where prediction_length > context_length.
TODO:
[x] Update docstrings
[x] Test the training script
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.
In general, I don't disagree that names can be improved but I am wondering if encode and decode have other connotations, i.e., encoder and decoder in an encoder-decoder transformer model.
Description of changes: This splits
input_transform
intocontext_input_transform
andlabel_input_transform
. Previously,input_transform
was being used for both context and label during training which would lead to incorrect results whereprediction_length
>context_length
.TODO:
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.