BENDR: Using Transformers and a Contrastive Self-Supervised Learning Task to Learn From Massive Amounts of EEG Data

脳波から自己教師あり学習する研究すでにあったすげ〜 https://t.co/F6986N1A0s

RT @AV_SP: Interesting idea-using transformer-based DNNs to model large EEG data sets: specifically adapting wav2vec2 - to get at general E…

Interesting idea-using transformer-based DNNs to model large EEG data sets: specifically adapting wav2vec2 - to get at general EEG 'features' https://t.co/oUJDcVfXAU Github https://t.co/Jxx1ST1Fn0 - attention needs lots of memory & compute - maybe chec