Is it possible to use only the encoder part of a Seq2Seq model ? I want to do some data analysis on the encoding after training
Similarly for the decoder portion. I want to see what gets generated by inputting random vector inputs
Is this possible using HuggingFace?
Thank you!