Wav2Vec2: how much context is available for self-attention

Models
1

Hi all, thanks as always for this great resource. I’ve essentially deleted this question because it has evolved into more of a newbie Pytorch question. Sorry for posting.