search

weiyinwei / MMGCN

MMGCN: Multi-modal Graph Convolution Network forPersonalized Recommendation of Micro-video
280 stars 52 forks source link

movielens dataset collect and process #57

Open PCLttb opened 9 months ago

PCLttb commented 9 months ago

Hi @weiyinwei , I appreciate your work. I tried to collect MovieLens dataset, and I have a question:

In terms of acoustic modality, I adopt VGGish to learn the acoustic deep learning features . But the feature size learned by each audio is (C, 128), how does your paper deal with the dimension as (1, 128) ?

weiyinwei commented 8 months ago

\sum_{c=1}^C {vector_audio} / C