Self-Sueprvised Pretraining for Visual and Language Transformer using Contextual and Weakly-annotated Youtube dataset in this work I describe my method for self-supervised learning for cross-modality transformers and my method for obtaining large datasets using videos
Мохаммад А. (науч. рук. Фильченков А.А.) Self-Sueprvised Pretraining for Visual and Language Transformer using Contextual and Weakly-annotated Youtube data // Сборник тезисов докладов конгресса молодых ученых. Электронное издание. – СПб: Университет ИТМО, [2023]. URL: https://kmu.itmo.ru/digests/article/10193