doc-doc / HQGA

Video as Conditional Graph Hierarchy for Multi-Granular Question Answering (AAAI'22, Oral)
MIT License
31 stars 4 forks source link

Features of MSRVTT-QA #15

Open songbo125 opened 1 year ago

songbo125 commented 1 year ago

Thank you for sharing such great work! I am very interested in it! But I have many difficulties in MSRVTT feature extraction, and my GPU is relatively poor. Therefore, I wonder if it is convenient for you to share the video features and BERT features extracted on MSRVTT-QA, similar to the MSVD-QA you provided. This is my email: 1029678831@qq.com, Thanks a lot !!!

doc-doc commented 1 year ago

Hi, please find the feature here