How to extract vggish features having overlapping?

v-iashin / MDVC

PyTorch implementation of Multi-modal Dense Video Captioning (CVPR 2020 Workshops)

https://v-iashin.github.io/mdvc

143 stars 20 forks source link

How to extract vggish features having overlapping? #27

Open ttgeng233 opened 2 years ago

ttgeng233 commented 2 years ago

Hello, if I want to extract audio features that per feature is 0.96s long with 0.32s overlapping, can I just set the parameter "EXAMPLE_HOP_SECONDS=0.32" without any other changes in ./models/vggish/vggish_src/vggish_params.py? Are there any other changes required?

v-iashin commented 2 years ago

Hi, I think you and I are on the same page on this matter. I am afraid you have to inspect it on your own. However, please let us know the answer once you figure it out.

v-iashin commented 2 years ago

@Gttgithub did you figure it out?